A New Pipelined Architecture for JPEG2000 MQ-Coder
Abstract— JPEG2000 has become one of the most rewarding image coding standards. It provides a practical set of features which weren’t necessarily available in the previous still image coding standards. The features were realized as a result of two new techniques adopted in this standard, namely the Discrete Wavelet Transform (DWT), and Embedded Block Coding with Optimized Truncation (EBCOT). The generated coefficients by DWT are entropy coded by EBCOT algorithm. EBCOT is a two-tiered coder, where Tier-1 is a context-based adaptive arithmetic coder, and Tier-2 is a rate control algorithm. The complexity of EBCOT Tier-1 makes its hardware implementations very difficult. A high speed hardware implementation usually takes a large amount of space on the die.
In this paper we propose a new simplified pipelined architecture for the JPEG2000 MQ-Coder. The proposed approach has resulted in a 20% decrease in hardware requirements and 10% increase in clock frequency. Post synthesis simulations indicate that the proposed architecture is able to compress 4 CIF video (704×576 pixels) at a rate of 30 frames per second, making it a good candidate for high resolution real time video coding, or high speed compression of high resolution images.
Index Terms— Byte-out, CODELPS, CODEMPS, EBCOT, flush, JPEG2000, MQ-Coder, Renormalization, Tile-Component
I.INTRODUCTION
mage data compression has always been a necessary and prominent issue due to boundaries of data bandwidth and storage. JPEG [1][2], a traditional standard of coding, has proved to be a suitable technique for compressing natural images at high bit rates. Yet the imperfections due to the blocking effect make this technique impractical especially for low bit rates of image compression.
JPEG2000 [3]-[8] has recently been proposed as a new high performance and multi-featured, yet complex standard of digital image coding. JPEG2000 offers numerous advantages over JPEG. These advantages include: ROI (Region Of Interest) coding, quality vs. resolution compression, lossless and lossy compression, progressive image compression/transmission by resolution/quality, random code-stream access and error resilience. Such characteristics add to the functionality of a system that is M. Ahmadvand is with the Department of Computer Engineering & IT, Hamedan University of Technology, Hamedan, Iran (e-mail: ahmadvand@hut.ac.ir).
A. Ezhdehakosh is with Department of Computer Engineering & IT, Amirkabir University of Technology, Tehran, Iran (e-mail: ezhdehakosh@aut.ac.ir). employing JPEG2000 as an image compression technique. The features and performance of JPEG2000 make this standard superior to JPEG. Yet computational complexities of JPEG2000 are much higher than that of JPEG. Such complexities are due to EBCOT [9][10] as the most important algorithm employed in JPEG2000. That is why EBCOT algorithm plays a major role in hardware implementation of JPEG2000 in different applications. During the process of encoding, an image is partitioned into data matrices called Tile-components. Each Tile-component is then coded separately. The process of coding is made up of different sections. These sections are depicted in Figure 1 and each is described below.
Fig. 1. JPEG2000 encoder block diagram
https://www.sodocs.net/doc/bc1281702.html,ponent Transform
This section is optional in JPEG2000 and is used to improve compression efficiency [11]. The transform converts the RGB data into another color representation, with a luminance (or intensity) channel and two color difference channels. This is used for taking advantage of some of the redundancy between the original RGB components. In particular color difference components mostly account for less than 20% of the bits used to compress a color image; therefore they are better represented as individual components [4].
B.Discrete Wavelet Transform (DWT)
DWT [4] is a domain transform that transforms an image Tile-component from special domain to frequency domain and provides a special decorrelation. This transform can be executed for as many levels as necessary. The output of each level of DWT is categorized into four sub-bands. Each sub-band contains the high/low frequency characteristics of the input image.
C.Quantization
Quantization [3] is the process by which the sub-band samples generated by the DWT are mapped onto quantization indices for coding. This process is lossy unless the quantization step is one and the coefficients are integer. D.EBCOT Tier-1
This section receives the quantized wavelet coefficients and encodes them into bit-streams. These coefficients are sliced into code-blocks before they are fed into the EBCOT Tier-1 [9][10]. EBCOT Tier-1 is composed of two parts:
A New Pipelined Architecture for JPEG2000
MQ-Coder
M. Ahmadvand, and A. Ezhdehakosh
I
Bit-Modeler and MQ-Coder [4]. Bit-Modeler is a bit-plane (a matrix that contains all the bits of the same order of all the coefficients of a code block) coder. A Bit-Modeler exploits the symmetries and redundancies within and across the bit-planes and generates corresponding contexts for each bit. After the context is generated, the MQ-Coder will code the bits (decisions) based on their associated contexts. The MQ-Coder is a derivative of Q-Coder [12] and generates compressed bit-streams for every code-block. The detail functionality of MQ-Coder is explained in Section II. E. EBCOT Tier-2
This tier is for rate allocation. The rate allocation is responsible for acquiring the highest quality for the output while maintaining a predetermined resolution, or acquiring the highest resolution while maintaining a predetermined image quality. At EBCOT Tier-2 [3] the bit-streams generated by the Tier-1 is collected with their rate-distortion information. Then different truncation points are set according to the optimization diagram. Each truncation point determines how many bits of a relevant bit-stream are to be selected for the final bit-stream.
TABLE I
R UN T IME P ERCENTAGE OF D IFFERENT M ODULES IN JPEG2000 E NCODER
Operation
Lossy Lossless Component Transform 10.1 3.64 DWT
25.14 10.41 Quantization 6.4 N.A. EBCOT Tier-1 44.86 67.35 EBCOT Tier-2
13.5 18.6
The execution time of different modules in the JPEG2000 algorithm is presented in Table I. It is noted from this table that EBCOT algorithm, as one of the main modules in JPEG2000 standard, occupies over half of the execution time of the whole procedure. It is also noted that the complexity weight of EBCOT lies within Tier-1. Therefore,
the architectures proposed to reduce hardware resources of Tier-1 while maintaining a high throughput, are highly valued.
In this paper we propose a novel pipelined architecture for JPEG2000 MQ-Coder. In our proposed architecture we have focused on reducing the hardware resources while securing a high throughput for the design. Our main contribution in this design is a special trade-off between area and execution time.
This paper is organized at follows: in the next section a deep analysis of the MQ-Coder will be presented. In section III the proposed architecture for the MQ-Coder are reviewed. Our proposed architecture is presented at section IV. Synthesis results are depicted in section V followed by conclusions and references.
II. MQ-C ODER A LGORITHMS AND A NALYSIS
MQ-Coder is a module applied in JPEG2000 EBCOT
Tier-1 for generating output bit-streams [3][4][9]. MQ-Coder is an adaptive Binary Arithmetic Coder (BAC). The
functionality of BAC is discussed in the following sub-section.
A. Binary Arithmetic Coder (BAC)
In BAC [13], symbols (either logic '0' or logic '1') in a code-stream are classified as either More Probable Symbol (MPS) or Less Probable Symbol (LPS) [13]. The probability of the occurrence of the MPS is called Pe and the probability of the occurrence of LPS is called Qe. Either of the symbols 0 or 1 can be MPS or LPS depending on the probability of their occurrence. In BAC an interval is considered in order to represent the probability of MPS and LPS. The initial interval is [0,1) and is divided to subintervals corresponding to the values of Qe and Pe. When a symbol occurs (either MPS or LPS), the subinterval associated with that symbol becomes the new interval. When the last symbol has been received a code-word C will be developed. The code-word C always points to the left point (lower bound) of the interval and A denotes its width. The BAC algorithm needs multiplication for the coding of each symbol, which is an area and time consuming operation for hardware implementation. Also, since a compressed data will only be generated when the last symbol of an input stream has been received by the encoder, an implementation of this algorithm will be exposed to serious loss of data at the times that the last bit of a stream is not received. Finally, after each update the length of the code-word and interval will be grown. This leads to a need for a high number of bits for storing the code-word and interval in hardware implementation of the algorithm. A specific type of efficient BAC which has been adapted to deal with the issues discussed above has been developed and is called MQ-Coder.
B. MQ-Coder
This adaptive binary arithmetic coder is used in JPEG2000 standard. In order to omit the multiplication operations, the length of interval A is maintained in the range [0.75,1.5). This means that the interval is always approximately equal to 1, if rounding to one significant bit. Therefore Qe ×A ≈ Qe and the following changes will be made to step 3 of the encoding process of BAC. MPS occurrence: C = C + Qe and A = A – Qe. LPS occurrence: A = Qe. The value C is kept in a 32 bit code-word [3][4] as shown in Figure 2.
In MQ-Coder, the last byte of the code-word is being sent to the output at special times, therefore the problems associated with the growing length of a code-word and compressed data being generated only after receiving the code-word's last bit is removed. The MQ-Coder can be understood as a module illustrated in Figure 3, which maps a sequence of input symbols (decisions) and associated contexts, to a single compressed code-word. The MQ-Coder utilizes a probability model for its
encoding process. This model is implemented as a Finite
State-Machine (FSM) of 47 states. In this state-machine,
each state contains coding information. Coding information
determines whether the current MPS (Most Probable
Symbol) is 0 or 1. If an MPS has occurred the CODEMPS
algorithm is performed, while if an LPS has occurred the CODELPS algorithm is performed.
1.CODEMPS Algorithm
If an MPS has occurred, the CODEMPS procedure [3] is called. The length of interval A is updated to (A – Qe) and code-word C is updated to (C + Qe). The value of A is always checked after it has been updated to determine if it has fallen below 0.75. If it does, it could mean that (A – Qe) has fallen below the value of Qe meaning that the subinterval associated with MPS is smaller than the subinterval associated with LPS. Therefore the two subintervals must be changed.
2.CODELPS Algorithm
If an LPS has occurred, the CODELPS procedure [3] is called. The length of the interval A is updated to value Qe, while the code-word C remains unchanged. If the LPS occurs successively for many times, Qe would become progressively larger and eventually (A – Qe) would become less than Qe. Therefore the portion of interval A that represents probability associated with LPS would become larger than the portion representing the probability of MPS. However, this does not occur since the CODELPS procedure tests for this condition and swaps the intervals associated with LPS and MPS when necessary.
3.Renormalization Algorithm
In order to ensure that the interval value A always remains in the range of [0.75,1.5), a renormalization [3] method is applied. The value A would fall below the value of 0.75 at the times that so many MPS has occurred. This case is also true for every time that an LPS occurs. This is due the fact that the value Qe, which interval A is updated to, is always less than 0.75. The renormalization algorithm shifts the values of A and C every time it is applied. The value of C code-word is sent to the output as compressed data by the byte-out and flush algorithms. These algorithms are discussed in the following sub-sections.
4.Byte-out Algorithm
The byte-out algorithm [3] generates the current byte-out value regarding the value of the last byte-out and carry bit in the code-word. As it was mentioned before the value of C is added with the Qe value each time a new decision has been received and value of C has been changed. If a carry bit (bit 'c' in C code-word) is generated from this addition, it must be added to the last generated byte-out. If the last byte-out becomes 0xFF, the carry bit is sent individually along with the last byte-out in order to prevent further carry propagation. This is called bit-stuffing.
5.Flush Algorithm
The flush algorithm is composed of different parts. At first a set-bit algorithm is performed in order to detect the best value of C, so that the lower two bytes of C contains 16
or 15 bits with the value of 1. After the set-bit algorithm [3], the byte-out algorithm is called for two times. At the end, the last compressed data is checked whether it is 0xFF or not. If it is, nothing is sent to the output since the 0xFF are not sent as the encoded data.
The complexity of the different algorithms in MQ-Coder, make its implementation very difficult. Some challenges encountered in the implementation of these algorithms are as follows:
1.The encoding procedures are serial processes with
high dependency. Therefore it is impossible to
employ parallel processing in the implementation
of these procedures.
2.Renormalization is a time consuming process, which
is due to the sequential shifts that are employed in
this algorithm. Therefore a novel technique for
reducing the execution time of this process is
necessary for hardware implementation.
3.A large number of calculations are applied for each
context-decision input. This leads to a long
execution time and too many resources in hardware
implementation.
In order to achieve a good hardware implementation of the MQ-Coder, one must take the aforementioned challenges into consideration.
III.P REVIOUSLY P ROPOSED A RCHITECTURES
In this section the hardware implementation of different architectures are compared. Several MQ-Coder implementations have been introduced in the literature. These implementations were using either pipelined or non-pipelined architectures.
Non-pipelined architectures [14] suffer from a low clock frequency and a very low throughput. Several buffers are required for interfacing such architectures with the rest of the components and in order to match their low throughput. This not only affects the overall performance but also increases the hardware resource requirements.
Pipelined architectures consist of a sequence of pipeline stages. The architecture proposed in [15] consists of three stages. The first stage calculates the new interval and code-word results. In order to perform these operations the Qe value and other necessary information such as NMPS will be derived from FSM [3][4]. The renormalization algorithm
is also performed in this stage. The second stage is dedicated to the bit-stuffing algorithm. At the third stage four FIFO modules are used. The intricacy of the first stage leads to a critical path that affects the clock frequency. In this design there is no mechanism to prevent data hazards. Since data hazards occur quite often during an adaptive binary arithmetic coding process, the proposed architecture suffers from a large number of pipeline stalls. The flush algorithm is not supported in this design.
The proposed architecture in [16][17] has four pipeline stages which are more balanced compared to the previous architecture. The index and MPS sense of each context is stored at the first stage. At stage two a probability estimation model is implemented as a look-up table. At the third stage calculations for updating the value of interval is
performed. The renormalization algorithm is also fulfilled in this stage. Each shift applied in the renormalization algorithm takes one clock cycle in this design. The calculation and renormalization of code-word C is applied in the last pipeline stage. The compressed data is issued from this stage to the output.
In this design, a multi-port memory is used. This leads to an implementation that occupies a large area, and has a slow access time. Besides, as the renormalization algorithm is not implemented with a barrel shifter, there is more chance for pipeline stalls. The flush algorithm is not supported in this design either.
The proposed architecture proposed in [18] is composed of four pipeline stages. At the first stage the probability model is implemented and the state of each context is stored. The main function of stage two is to update the interval value. The renormalization algorithm is also performed in this stage. The update and renormalization of the 16 lower bits of the C code-word, is done in this stage. The rest of the bits in C is updated and renormalized at the last stage. This is done in order to shorten the critical path. The flush and byte-out algorithms are also performed at the last stage.
This design suffers from the slow access time and large area consumption caused by employing a multi-port memory. The renormalization algorithm is not implemented with a barrel shifter, which adds to the chance of receiving pipeline stalls.
The architecture proposed in [19] is a five pipelined stage. The states of contexts are stored at the first stage. The probability model is implemented at stage. The update and renormalization of interval value is performed at the third stage. The code-word C is updated and renormalized at the fourth stage. At the last stage the byte-out and bit-stuffing algorithms are implemented.
The design has balanced stages and no stalls. But this has consumed a lot of hardware resources which in turn has lead to large area consumption. The complexity of this design is mostly caused by the attempt to eliminate pipeline stalls. The flush algorithm is not supported in this design either. The observed imperfections in the proposed architectures were: unbalanced pipeline stages, not supporting all the algorithms present in the MQ-Coder, lack of solution for removing data hazards, high area consumption and inefficient implementation for the renormalization algorithm. We must note that each one of the proposed architectures has some of the defects mentioned above.
IV.P ROPOSED A RCHITECTURE
Our proposed architecture consists of five pipeline stages as shown in Figure 4. These stages are as follows: 1) Context-Decision Fetch with data-Forwarding (CDFF), 2) Probability Estimation (PE), 3) Interval Update (IU), 4) Code-word Update (CU) and 5) Byte-Out (BO). Each stage is described in details below.
A.Context-Decision Fetch with data-Forwarding (CDFF) This is the first stage (Figure 4.a.) of our pipelined architecture. The inputs of this stage are a 5-bit context value (ctx-read) along with a single bit decision signal (decision). The duty of this stage is to generate the state of input value ctx-read. The state includes index of Qe table (look-up table that represents the probability model) and MPS sense [4]. This is done through a Context-Table module that contains a state value for every context. These state values are updated through a feedback context called ctx-write. The current state of the signal ctx-write is updated to new-state which is provided by the next stages. In case that two similar context values are fed to this stage consecutively, the new-state signal is saved and sent to the state output of this stage directly, thus avoiding pipelined stalls. Detecting such cases is done by a Data-Forward-detector module that compares the values of two consecutive contexts.
B.Probability Estimation (PE)
The main module of this stage (Figure 4.b.) is Qe-Table [3] which in fact represents the probability model. Each entry of this table contains probability value (Qe), new index in the case of MPS or LPS occurrence (NMPS and NLPS) and switch (SW) signal [3]. The normal practice in pipeline architectures is to employ a multi-port memory in order to implement the Qe-table. However, our design utilizes a special technique in order to replace the multi-port memory with a single port memory. In this technique the NMPS and NLPS value corresponding to the last context are always stored. In order to handle the data-hazard situation, the index of the current context or the NMPS or the NLPS of the last context is used as the correct index of the current context. It must be noted that cases in which data-hazard occurs are detected by the data-forward-detector in the previous stage. Whether the index of the current context or the NMPS/NLPS of the last context is selected, is done by a module named index-selector. By utilizing this technique the design has removed the pipeline stalls of this stage with minimum hardware overhead.
C.Interval Update (IU)
The main task of this stage (Figure 4.c.) is to update the interval length (A) of MQ-Coder. This interval is stored in a 16-bit register called A-Register. In order to update the A value to A – Qe, a 16-bit subtractor is employed. The output of the subtractor is multiplexed with Qe-value and is passed to Zero-Detector and A-Barrel-Shifter as a new interval length. The Zero-Detector module detects the number of consecutive zero bits at most significant positions of this new interval length (Zero-Num). The Zero-Num value is the number of shifts required for the renormalization algorithm and is passed to A-Barrel-Shifter and to the next stage. The new interval length is shifted by the A-Barrel-Shifter in a single clock cycle. The shifted value is stored in A-Register as the renormalized interval length. The normal practice in pipelined architectures is to implement hardware for a maximum of 15 shifts and dedicate one clock cycle per each shift. This results in pipeline stalls equal to the number of shifts. In our architecture the shift is performed by A-Barrel-Shifter and therefore avoiding extra clock cycles. Besides, the renormalization of the A interval is perform ed for 79% (as shown in Figure 5) of the times and
(b) PE
(c) IU
(d) CU
(e) BO
Context Decision
(a) CDFF
Fig. 4. Proposed pipelined architecture data path
Fig. 5. Renormalization occurrence probability
the maximum number of shifts applied for each renormalization is equal to 15, yet as the simulations results presented in Figure 6 indicate, the number of shifts is less than 8 for more than 90% of the time. In our design in order to reduce hardware resources and increase clock frequency, we propose that a maximum of 7 shifts per clock cycle be implemented. A Hold-State module is employed in order to extend the shifting operations for one more clock cycle at the times that the number of shifts is more than 7. This will result in one clock stall in 7.9% of condition in total.
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
Statistical Samples
Fig. 6. Number of shifts for various renormalizations Another module in this stage is the Update-State unit which is used to update new state of current context and also determines if decision is MPS or LPS. The outputs of this module are sent to the context-table of stage 1 as ctx-write and new-state.
A.Code-word Update (CU)
The updating of code-word value is performed in this stage (Figure 4.d.). The code-word value (C) is stored in a 32-bit register called C-Register. The number of shifts that should be performed over the A value and C code-word before every byte-out is stored in a 4-bit register called CT-Register. It should be mentioned that every time the CT-Register becomes zero a byte-out is sent to the next stage. After every byte-out this register is initialized to the value determined by the Init-CT. In order to update the CT-Register a 4-bit subtractor called A-subtractor is used to reduce the CT-Register by Zero-Num. Normally the output of the A-subtractor is positive. Yet in some occasions when CT-Register is less than Zero-Num the result becomes negative. In order to correct this, a 4-bit adder called Init-Adder is employed.
In order to generate the new code-word and therefore perform the renormalization algorithm, the C + Qe value and the current code-word are multiplexed and the result is passed to the C-Barrel-Shifter. The number of shifts applied in the C-Barrel-Shifter is equal to the number of shifts applied to the A-value in the previous stage.
In the standard the shift operation for the code-word value is introduced so that each shift is performed at every clock cycle. The byte-out data is generated when enough shifts have been performed over the C code-word. The rest of the shift operations are applied after byte-out generation. However, since the shift operations are performed with a barrel shifter in our proposed architecture, the byte-out is produced after all the shifts are applied together. A C-Reg-Correction module is employed in order to recognize the correct location of the byte-out data in the C code-word.
The flush algorithm [3] is performed parallel to the updating and renormalization of the code-word. This algorithm is responsible for sending the last value of C-Register to the output in the end of coding. The suitable C value for best compression is the one that contains the most number of bits with the value 1. The Carry-Detector is employed in order to choose the suitable C value. This value is kept in a register called C-Flush-Reg. When the best value for C is stored, it must be shifted out in order to generate compressed data. This is done by the Flush-Shifter module.
B.Byte-Out (BO)
This stage (Figure 4.e.) performs the byte-out and Bit-stuffing algorithm [3]. An 8-bit register, B-Register, is used to store the last byte of compressed data. The last byte-out for the next clock is selected from the C-Register/C-Flush-Reg by the Byte-Selector module.
In order to implement the Bit-stuffing algorithm two modules are used: The All-One-Detector that determines whether the last byte-out is 0xFF or not and the B-Adder that adds the B–Register value with the value determined by All-One-Detector, namely the carry bit (27th bit of the code-word). The output of the B-Adder is sent to the output of this stage as the compressed data.
V.E XPERIMENTAL R ESULTS
The proposed architecture of MQ-Coder has been simulated using VHDL. This architecture is implemented by 0.18 μm CMOS technology. The synthesis results of the proposed architecture are shown in Table II. The gate count and clock frequency of this architecture is compared to three previous pipeline architectures with the same technology. The execution time of our proposed architecture is compared to other architectures at Table III using Lena, Baboon and Jet images with the resolution of 256 × 256 and 24-bit RGB components.
As the results indicate, the design in [15] suffers from a low clock frequency and consumes a large number of clock cycles during the coding process, resulting in a very low throughput. The design in [16] has a high clock frequency and a low gate count. Yet the number of clock cycles consumed for the coding process is very large, which has resulted in a low throughput as well.
TABLE II
C OMPARISON OF I MPLEMENTATION R ESULT
[15][16][19]Proposed Gate# 8459 7100 9156 7325 Clock Frequency
(MHz)
185.5 206.2 190.6 208.1
The architecture in [18][19] does not have a high clock frequency, but since the number of clock cycles for the encoding process is very low, therefore the coding time is acceptable. The only shortcoming of this design is its high number of gate count. In our proposed architecture although the number of clock cycles for an encoding process is higher than the architecture in [19], but due to its high clock frequency it has the lowest coding time. In addition the gate count of our proposed architecture is 20% lower than the next fastest design.
TABLE III
E XECUTION T IME FOR T HREE P ICTURES
Timing Lena Baboon Jet
[15]CLK # 1311233 1554195 1286333 Time (ms) 7.1 8.38 6.94
[16]CLK # 1297355 1605745 1276743 Time (ms) 6.29 7.79 6.19
[19]CLK # 998967 1224958 959033 Time (ms) 5.24 6.42 5.03
Our CLK # 1077885 1321729 1034796 Time (ms) 5.18 6.35 4.98
Post synthesis simulations indicate that the proposed architecture encodes precisely one context-decision pair every 1.079 clock cycle and operates at 208 MHz. This architecture is able to compress 4 CIF video (704×576 pixels) at a rate of 30 frames per second, making it a good candidate for high resolution real time video coding, or high speed compression of high resolution images.
VI.C ONCLUSION
A high-speed pipelined architecture with reduced area for JPEG2000 MQ-Coder is proposed in this paper. In this design the time consuming algorithms are divided into different pipeline stages. Therefore the critical path has been reduced considerably. All of the algorithms introduced in the JPEG2000 MQ-Coder are supported in this design. Special techniques employed in order to implement the renormalization algorithm, has led to major reduction in hardware recourse requirements and improving the clock frequency while receiving a few pipeline stalls. The stalls occur in 7.9% conditions in total. Therefore every context-decision pair is encoded in 1.079 clock cycle.
The architecture is implemented by 0.18 μm CMOS technology and is functional at 208 MHz clock frequency. This architecture being able to code 4 CIF video (704 × 567) at a rate of 30 frames per second is suitable for real time image processing applications.
R EFERENCES
[1]"Information Technology—JPEG—Digital Compression and Coding
of Continuous-Tone Still Image—Part 1: Requirement and Guidelines", ISO/IEC 10918-1 and ITU-T Recommendation T.81, (1994).
[2]W. B. Pennebaker and J. L. Mitchell, "JPEG Still Image Data
Compression Standard", New York: Van Nostrand Reinhold, (1992). [3]"JPEG2000 part I final draft international standard," ISO/IEC
JTC1/SC29/WG1 N1890, (2000).
[4] D. S. Taubman and M. W. Marcellin, JPEG2000: image compression
fundamentals, standards and practice, MA: Kluwer, Norwell, (2002). [5]M. D. Adams, H. Man, F. Kossentini, and T. Ebrahimi, “JPEG 2000:
The next generation still image compression standard,” Doc. ISO/IEC JTC1/SC29/WG1 N1734.
[6]Skodras, C. Christopoulos, and T. Ebrahimi, “The JPEG 2000 still
image compression standard,” IEEE Signal Processing Magazine, (2001).
[7]M. J. Gormish, D. Lee, M. W. Marcellin, “JPEG 2000: overview,
architecture, and applications,” in Proc. IEEE Int. Conf. Image Processing, 2, (2000).
[8]M. W. Marcellin, M. J. Gormish, A. Bilgin, and M. P. Boliek, “An
overview of JPEG-2000,” in Proc. IEEE Data Compression Conf.
(DCC2000), (2000).
[9] D. Taubman, "High performance scalable image compression with
EBCOT", IEEE Transaction on Image Processing, 9, 7, (2000).
[10] D. Taubman, E. Ordentlich, M. Weinberger, G. Seroussi, “Embedded
block coding in JPEG 2000,” Signal Processing: Image Communication, 17, (2002).
[11]K.-F. Chen, C.-J. Lian, H.-H. Chen, and L.-G. Chen, “Analysis and
architecture design of EBCOT in JPEG2000,” in Proc. IEEE Int.
Symp. Circuits and Systems (ISCAS’01), (2001).
[12] B. Pemmebaker, J. Mitchell, G. Langdon, R. Arps, "An Overview of
the Basic Principles of the Q-Coder Adaptive Binary Arithmetic Coder", IBM J. RES. DEVELOP, 32, (1988).
[13]G. G. Langdon Jr., "An Introduction to Arithmetic Coding", IBM
Journal of Research and Development, 28, (1984).
[14]K. Andra, C. Chakrabarti, T. Acharya, "A High-Performance
JPEG2000 Architecture", IEEE Trans. On Circuits and Systems for Video Technology, (2003).
[15]K.K. Ong, W.H. Cahng, Y.C. Tseng, Y.S. Lee, C.Y. Lee, "A High
Throghput Context-Based Adaptive Arithmetic Coder for JPEG2000", IEEE International Symposium on Circuits and Systems, (2007). [16]M. Tarui, M. Oshita, T. Onoye, I. Shirakawa, "High-Speed
Implementation of JBIG Arithmetic Coder", Proceedings of the IEEE Region 10 Conference, (2001).
[17]JBIG Bi-Level Image Compression Standard, ISO/IEC 11544 and
ITU-T Recommendation T.82, (2000).
[18] C. Lian, K. Chen, H. Chen, L. Chen, "Analysis and Architecture
Design of Block Coding Engine for EBCOT in JPEG2000", IEEE Transaction on Circuits and Systems for Video Technology, 13, (2003).
[19]M. Ahmadvand, A. Shahrokhi, O. Fatemi, "A High-Speed Pipelined
Architecture for MQ-Coder of JPEG2000 Standard", 27th Queen's Biennial Symposium on communications, (2009).
英语作文关于共享单车的篇精编
(一) 假定你是红星中学初三学生李华。你的美国朋友Jim在给你的邮件中提到他对中国新近出现的一种共享单车“mobike”很感兴趣，并请你做个简要介绍。请你给Jim回信，内容包括： 1. 这种单车的使用方法(如：APP查看车辆、扫码开锁等); 2. 这种单车的优势; 3. 你对这种单车的看法。 注意：1. 词数不少于80; 2. 开头和结尾已给出，不计入总词数。 提示词：智能手机smartphone, 二维码the QR code 参考范文 Dear Jim, I’m writing to tell you more about the new form of sharing bike mobike mentioned in your latest letter. It’s very convenient to use if you have a smartphone. What you do is find a nearest mobikethrough the APP, scan the QR code on the bike, and enjoy your trip. Compared to other forms of sharing bike, the greatest advantage of mobike is that you can easily find one and never worry about where to park it. It is becoming a new trend as a means of transportation, which relieves the traffic pressure and does good to the environment as well. Hope to ride a mobike with you in China. Yours, Li Hua (二) 最近很多大城市都投放了共享单车(shared bikes)，比如摩拜单车(Mobike)、Ofo共享单车等。由于它们方便停放，骑车也能起到锻炼身体的作用，作为代步工具很受大家欢迎。但是，各地也出现了很多毁车现象，比如刮掉车上的二维码(QR code)、上私锁等。 你对这种现象怎么看?你对共享单车公司有什么建议吗?写一篇符合逻辑的英语短文，80词左右。 参考词汇：bike-sharing companies 共享单车公司，Mobike 和Ofo 是两家共享单车公司，convenience 方便，register登记 参考范文 The shared bikes like Mobike and Ofo bring great convenience to people. You needn’t lock them by simply using your smart phone. They can take you where the subway and bus don’t go. And they can be left anywhere in public for the next user. However, bad things happen. Some people damage the QR code on the bike, or use their own lock, which causes trouble to other users. In my opinion, it’s difficult to turn these people’s ideas in a short time. Therefore, bike-sharing companies like Mobike and Ofo need to do something. For example, those who damage the bike should pay for their actions. Also, because people use their real name toregister as a user, it’s a good way to connect to one’s personal credit. In the end, what I want to say is to take good care of public services. (三) 共享单车(bicycle sharing)已成为时下最热的话题之一，请你就这一话题写一篇短文。内容须包括三方面：1. 共享单车蓬勃发展，成为社会热潮;2. 共享单车带来便利，但也存在问题;3. 我对解决问题的建议。 参考范文 Bicycle Sharing With the development of technology, bicycle sharing comes into people's lives. It becomes more and more popular and much news reported it. At the same time, we should see that there are some problems caused by bicycle sharing. On one side, bicycle sharing makes it very convenient of people traveling. You can find a bicycle anywhere at any time when you want to go out for a cycling, and the price of one trip is very low. It can save time for people. On the other side, its management is not perfect. Even kids can open the lock and ride the bicycle, there is no doubt that such behavior is very dangerous.
介绍北京的英语作文(2)
介绍北京的英语作文(2) AsBeijinghasbeenconfirmedhomecityofOlympics2008,the spiritofgreenOlympics,scientificOlympicsandhumanizedOlymp icwillsurelybringmoreandmorechangestoBeijing,promotethed evelopmentofsportsandOlympicsinChinaaswellasintheworld,a ndstrengthenthefriendlycommunicationsbetweenChineseandf oreignpeople. 篇六：Beijing BeijingisthecapitalofPeoplesRepublicofChinaandthenation scentreforpolitics,economyandculture.Itenjoysalongandrichhis tory.Therearenumerousheritagesitesandwonderfulexamplesof ancientarchitecture,suchastheworld-famousGreatWall,theTem pleofHeavenandtheForbiddenCity. Besidessightseeingplaces,therearemanydeliciousfoodsuch asPekingducksandBeijingsnacks.Beijingisreallyagoodplacetotr avel. 篇七：Beijing AsthecaptainofChina，Beijinghasbeenthemostpopularcityofchina。SomoreandmorepeoplewanttovisitBeijing.
汽车利弊英语作文4篇
[标签:标题] 篇一：关于汽车的英语作文 好的 Nowadays, with the rapid improvement of people’s living standards, cars have become an indispensable part of people's lives,so that more and more people have a car of their own, especially in cities. It brings some benefits for us but also causes many problems at the same time. For one thing,there’s no doubt that cars provide much convenience for people to go where they want to quickly and easily. Especially on weekday,driving a car can save a lot of time for us to go to work.When some places are too far away from our home, driving our own car is also convenient, we can go wherever we want. However,for another, too many cars will lead to the pressure of public transport, a series of problems will appear.First of all,it will bring about more air pollution,a large amount of polluted air given off by cars do great harm to our health.What’s more, as the existing roads are not so wide for the increasing number of cars,undoubtedly,traffic jams will become more and more serious. Last but not least, cars also place burden on the public facilities in providing more parking lots. As far as I am concerned,everything has its advantages and disadvantages. It’s high time that effective action must be token to limit the ever growing number of cars, the government should take measures to control the air pollution from the cars. Some roads should be widened and more new roads should be constructed. Only in this way，will people benefit from the popularity of cars. 坏的 Nowadays, with the rapid improvement of people's living standards, cars have become an indispensable part of people's lives,so that more and more people have a car of their own, especially in cities.It brings some benefit for us but also causes many problems at the same time. For one thing,it's no doubt that that cars provide much convenience for people to go where they want to quickly and easily. Especially on weekday,driving a car can save a lot of time for us to go to work.When some places are too far away from our home, driving our own car is also convenient, we can go wherever we want. However,for another, too many cars will lead to the pressure of public transport, a series of problems will appear.First of all,it will bring about more air pollution,a large amount of polluted air given off by cars do great harm to our health .What's more, as the existing roads are not so wide for the increasing number of cars,undoubtedly,traffic jams will become more and more serious. Last but not least, cars also place burden on the public facilities in providing more parking lots. As far as I am concerned,everything has its advantages and disadvantages. It's high time that effective action must be token to limit the ever growing number of cars, the government should take measures to control the air pollution from the cars. Some roads should be widened and more new roads should be constructed. Only in this way,will people benefit from the popularity of cars. 篇二：雅思作文高分范文：私家车的利与弊 智课网IELTS备考资料 雅思作文高分范文：私家车的利与弊
2课下作业三十二
课下作业(三十二) 、选择题 1．哲学基本问题又称哲学的根本问题、哲学的最高问题。这一问题包括( ) ①物质和意识的辩证关系问题 ②思维和存在何者是本原的问题 ③思维和存在有没有同一性的问题 ④唯物主义和唯心主义关系问题 A ．①② B ．②③ C .③④ D .①③ 解析：选 B 。哲学基本问题包括两个方面的内容：一是思维和存在何者是本原的问题；二是思维和存在有没有同一性的问题，②③入选。 2. 唯物主义是哲学上两个敌对的基本派别之一，是同唯心主义相对立的思想体系。划分唯物 主义和唯心主义的唯一标准是( ) A .物质和意识的关系问题 B .客观与主观的关系问题 C .思维和存在何者是本原的问题 D ?思维和存在有没有同一性的问题 解析：选C。对思维和存在何者是本原问题的不同回答，是划分唯物主义和唯心主义的唯一 标准， C 入选。 3. “二月春分八月秋分昼夜不长不短；三年一闰五年再闰阴阳无差无错。”这副对联从一个侧面反映了( ) ①思维和存在具有同一性 ②认识与自然的吻合具有必然性 ③认识以实证和猜测为基础 ④意识活动具有主动创造性 A .①③ B .②④ C .①④ D .②③ 解析：选C o材料反映人们可以认识和把握自然界的运动规律,说明思维和存在具有同一性 反映了意识活动具有主动创造性，①④符合题意；认识与自然的吻合不具有必然性，②错误；实践是认识的基础，③错误。 4 .(2019河南中原名校联考)“为天地立心，为生民立命，为往圣继绝学，为万世开太平”是北宋张载的名言。由于其言简意宏，一直被人们传颂不衰。下列观点符合“为天地立心”的
是( ) ①形存则神存，形谢则神灭②吾心即是宇宙，宇宙即是吾心 ③思维着的精神是地球上最 美的花朵④“天不生仲尼，万古长如夜” A ．①② B ．①③ C .③④ D .②④ 解析：选C。“为天地立心”的意思是为天地确立一种核心价值理念，强调精神的作用，③ ④强调思维着的精神的作用，符合题意；①强调物质决定意识，②片面夸大意识的作用，均不合题意。 5.有位生物学学者认为，唯有生物学才能带领人类探究物种本源、生从何来死往何方等问题，他确信构成生物的眼见为实的物质比辩证唯物主义判断猜想的物质还正确。这种认识( ) ①没弄清辩证唯物主义的物质概念与构成生物的物质之间的关系 ②坚持了唯物主义根本方向，但属于古代朴素唯物主义的思想 ③犯了近代形而上学唯物主义错误，具有机械性、形而上学性 ④是在自然科学基础上对辩证唯物主义和历史唯物主义的发展 A .①② B .①③ C .②③ D .③④ 解析：选B。材料中的生物学学者没有看到哲学对具体科学的意义，认为哲学中的物质是猜 想的，没有看到哲学中的物质与自然科学中的物质是共性与个性的关系，因而具有机械性、 形而上学性，①③正确，②④错误。 6.18世纪法国哲学家丹尼斯狄德罗认为：“自然界由数目无穷、性质不同的异质元素构成。” 这种观点( ) ①承认世界的物质性，但把物质归结为自然科学意义上的元素②建立在自然科学成就的基 础上，丰富和发展了唯物主义③坚持物质第一性，但对物质的认识没有科学依据④认为世界是物质的，正确揭示了物质世界的基本规律 A .①② B .③④ C .①③ D .②④ 解析：选A。“狄德罗认为：‘自然界由数目无穷、性质不同的异质元素构成。'” 这种观点承认世界的物质性，但把物质归结为自然科学意义上的元素，①正确；狄德罗生活在18世纪，其观点建立在自然科学成就的基础上，丰富和发展了唯物主义，属于近代形而上学唯物主义，②入选；③④ 说法错误。 7 ?“宇宙创造过程中，上帝没有位置……没有必要借助上帝来为宇宙按下启动键”。这是斯
课后巩固作业(十九) 3.2.1
温馨提示： 此套题为Word版，请按住Ctrl,滑动鼠标滚轴，调节合适的观看比例，答案解析附后。 课后巩固作业（十九） (30分钟50分) 一、选择题(每题4分，共16分) 1.下列试验中，是古典概型的是( ) (A)发射一颗卫星能否成功 (B)从高一(18)班60名同学中任选一人测量其身高 (C)抛掷一枚骰子，出现1点或2点 (D)射击选手射击一次，恰中靶心 2.(2011·温州高一检测)抛掷两个骰子,则两个骰子点数之和大于4的概率为 ( ) (A)13 18 (B)8 9 (C)7 12 (D)5 6 3.从{1,2,3,4,5}中随机选取一个数为a，从{1,2,3}中随机选取一个数为b，则 b>a的概率是( ) (A)4 5 (B)3 5 (C) 2 5 (D)1 5 4.用1，2，3，4这四个数字，组成比2 000大且无重复数字的四位数的概率是 ( ) (A)1 4 (B)1 2 (C)3 4 (D)1 3 二、填空题(每题4分，共8分) 5.(2011·江苏高考)从1，2，3，4这四个数中一次随机取两个数，则其中一个
数是另一个的两倍的概率是_____. 6.(2011·永定高二检测)扔两枚骰子出现的点数为m,n，以(m,n)为坐标的点出现在x2+y2=16内的概率为_____. 三、解答题(每题8分，共16分) 7.一个袋中装有四个形状大小完全相同的球，球的编号分别为1，2，3，4. (1)从袋中随机抽取两个球，求取出的球的编号之和不大于4的概率； (2)先从袋中随机取一个球，该球的编号为m，将球放回袋中，然后再从袋中随机取一个球，该球的编号为n，求n 介绍北京的英语作文1篇 篇一MyFamily Ilovemyfamily,becauseIhaveahappyfamily. MyfatherisanEnglishteacher.HisnameisJacky.Heisthirty-eight.Helikesplay ingbasketball.What’smymotherjop?Issheateacher?Yes,you’reright!Mymotherisverykindandnice,sheisthirty-seven.Mymotherisalways laboriouswork.Ilovemyparents! OnStaurdayandSunday,Ioftengotothelibraryandplaythepiano,Myfathergot oplaybasketball.Sometimes,wewatchTVandlistentomusicathome. Ilovemyfamily.BecauseI’mveryhappytolivewithmyparentstogether! 篇二MyFamily MyFamily Everyonehasafamily.Weliveinitandfeelverywarm.Therearethreepersonsin myfamily,mymother,fatherandI.Welivetogetherveryhappilyandtherearema nyinterestingstoriesaboutmyfamily. Myfatherisahard-workingman.Heworksasadoctor.Healwaystrieshisbesttoh elpevery,patientandmakepatientscomfortable.Butsonetimesheworkssohard thathecan”trememberthedate. 人教版二年级下册语文 每课作业 HEN system office room 【HEN16H-HENS2AHENS8Q8-HENH1688】 1、找春天 一、看拼音写词语。 tuō xiàjiědònɡxī shuǐmián yī tàn tóu yáo tóu?yě huā duǒcánɡ 二、按课文内容填空。 1、春天来了！我们几个孩子，_______棉袄，________家门，________田野，去_______春天。春天像个___________，遮遮掩掩，躲躲藏藏。我们________地找啊，找啊。 2、春天来了！我们______到了她，我们______到了她，我们______到了她，我们______到了她。她在柳枝上_________，在风筝尾巴上_________；她在喜鹊、杜鹃__________，在桃花、杏花__________…… 三、填空。 藏藏丁丁遮遮 慌慌清清摇摇 四、照样子写词语。 遮掩遮遮掩掩认真 开心（） （）（） 五、在正确读音下面画横线。 1、我家院子里有几（jǐ jī）只小鸡，这些小鸡真可爱！ 2、我家的客厅里的茶几（jǐ jī）上有两杯水。 ★照样子，写句子。 小草从地下探出头来，那是春天的眉毛吧？ ，是 2、古诗两首 一、看拼音写词语。 Wèi lái zhuī gǎn shūdiàn kūhuáng guāng róng yěcài zhù sù gāo shāo 二、抄写古诗《草》 题目： 作者：三、比一比，组词语。 未（）烧（）枯（）徐（） 末（）浇（）姑（）除（） ★这两首诗都是描写季景色的。 ★《草》是朝大诗人写的，诗中“， 。”句最为有名，它描写了草顽强的生命力。诗中还有一对反义词，分别是________和________。★《宿新市徐公店》描写了种景物，分别是 。诗中最有意思的一句 是、 。 3、笋芽儿 一、拼一拼，写一写。 shì jiè sǔn yá hūhuàn shān gāng hōng dòng 二、读一读，并照样子，写一写。 啊，多么明亮、多么美丽的世界呀！ 三、按课文内容填空。 1、笋芽儿被叫醒了。她（）眼睛，（）懒腰，（）四周一片漆黑，撒娇地说：“是谁在叫我呀” 2、春雨姑娘（）着她，（）着她。太阳公公（）着她，（）着她。笋芽儿脱下一件件衣服，长成了一株（）的竹子。她站在山冈上，（）地喊着：“我长大啦！”四、比一比，再组词。 岗（）界（）笋（）牙（）喊 （）呼（） 刚（）介（）笛（）芽（）减 （）乎（） 部编版三年级语文下册同步练习 20. 肥皂泡 一、看拼音写词语。 féi zào cháng láng mù wǎn tòu míng jiāo ruǎn yǎng tóu yí chuàn fēi yuè yīng ér xī wàng 二、给下列加点字选择正确的读音，打“√”。 1、几个同学散．（sǎn sàn）漫得在操场上散．（sǎn sàn）步。 2、若用扇．(shān shàn)子在下面轻轻地扇．(shān shàn)送,有时它能飞得很高很高。 三、比一比，再组词。 皂（）透（）婴（）娇（）希（） 泉（）诱（）耍（）骄（）杀（） 四、照样子，完成练习。 例：无声地散裂无声地（）无声地（） 例：轻轻地扇轻轻地（）轻轻地（） 例：颤巍巍（ABB式）（）（）（）（） 五、我会查字典，给加点字选择正确的解释。 “仰”字按部首查字法应先查（）部，再查（）画。按音序查字法应先查大写字母（），再查音节（）。“仰”字在字典里的解释有：①脸向上，与“俯”相对；②敬慕；③依赖；给下列加点字选择正确的解释，填序号。 仰．起头（）久仰．大名（）仰．仗（） 六、择合适的词语填空。（填序号） ①颤巍巍②软悠悠③轻飘飘④慢腾腾 1.不知为什么，我这几天头晕眼花，走起路来感觉________的。 2.轻轻地一提，那轻圆的球儿便从管上落了下来，_______地在空中飘游。 3.他做事总是像老牛拉破车，_______的。 4.有时吹得太大了，扇得太急了，这脆薄的球，会扯成长圆的形式，______的，光影零乱。 七、按要求完成句子练习。 1. 把句子写具体。 ①薄球就散裂了。___________________________________________________________ ②浮光在球面上乱转。______________________________________________________ 2.我把用剩的碎肥皂放在一只小木碗里。（改为“被”字句） ____________________________________________________________________________ 3.小时候，我喜欢玩吹泡泡、滚铁环、西瓜等游戏。（修改病句） ___________________________________________________________________________ 七、阅读课文片段,完成练习 方法是把用剩的碎肥皂放在一只小木碗里，加上点水，（），使它融化，然后 ．．用一 支竹笔的套管，蘸上那黏稠的肥皂水，（）地吹起，吹起一个（）的网球大小 的泡儿，再．（）地一提，那轻圆的球儿，便从管上落了下来，（）地在空中飘游。若用扇子在下面（）地扇送，有时它能飞得很高很高。 1.根据课文内容填空。 2.这是写（）肥皂泡和（）肥皂泡的过程。 3.用“”画出做肥皂水方法的句子。 4.用“先……然后……再……”说一说吹肥皂泡的过程。 . 八、快乐阅读。 种一片树叶 冰心 埋下一片树叶，固执地相信它终会长成一棵参天大树。这便是儿时最美的梦。 我4岁那年的秋天，枯黄的叶子从高高的树枝上飘落，正在院中玩耍的我俯身拾起一片， 觉得很美。玩赏了半天，我忽发一个奇怪的想法：这个曾经是树伯伯身上一部分的叶，把它 种进土里之后，必定可以长出一棵同样的大树! 于是我捡了许多片叶子，虔诚地跪在地上挖了个小坑，把树叶全埋到了土里。就像妈妈 生下了我，我也会渐渐长大—样，我坚定地相信我的梦马上可以实现。 我天天蹲在种树叶的地方等待我的大树，直到大雪纷飞。我想叶子大概怕冷，所以不敢 出来，等明年开春，一定会长得更好!我依旧固执地坚守着自己的梦，直到燕子衔泥，杨树吐 汽车的重要性《英语作文》 The automobile has become one of the most important means/ways of transportation in the world since it was invented. The automobile has completely changed the lifestyles of almost all the people in the world. In the past, animals like horses and camels were used for traveling and transporting goods. Automobiles are more comfortable and faster. Automobiles have also made it possible for us to transport large quantities of goods and people at the same time. Besides, the invention of the automobile has provided jobs for millions of people all over the world. 翻译： 汽车已经成为世界上最重要的交通工具之一，因为它是发明的。汽车已经完全改变了世界上几乎所有的人的生活方式。 在过去，像马和骆驼的动物被用来运送货物。汽车更舒适，更快速。汽车也使我们能够在同一时间运送大量货物和人。 此外，汽车的发明为全世界上百万的人提供了工作。 三一文库（https://www.sodocs.net/doc/bc1281702.html,） 〔我想去北京英语作文〕 我想去北京的英语作文如何写？那么，下面是小编给大家整理收集的我想去北京英语作文，供大家阅读参考。 我想去北京英语作文1 I’d like to go to a beautiful place. I think it would be Beijing. Beijing is not only our capital city, but also a famous city with long history and wonderful culture. Beijing is also China’s political and cultural center. There’re many old places of great interest, such as the Great Wall, the Summer Palace, the Forbidden City, the Temple of Heaven, and Tiananmen Square. Once you see Tiananmen Square, you will think of Beijing. It has been the symbol of Beijing since 1949. 我想去北京英语作文2 I went to Beijing more than eight times. Beijing is the capital of China. It’s a big city. I am very familiar with Beijing. It takes an hour and forty minutes from Nantong to Beijing by plane. There are many tall buildings in Beijing. It’s a modern city. My family visited the Great Wall, the Summer 必修2 Unit 1 课下作业（一~三） 课下作业（一）考点过关针对练 Ⅰ.单词拼写 1．Jim insisted that the book Mr.Black referred to was worth (值得的) reading. 2．She took one look at the horse and her heart sank (下沉)． 3．It was not until he removed (摘掉) his sunglasses that I recognized him. 4．I'll be waiting for you at the entrance (入口) to Guangzhou station. 5．The coal industry is now barely half its former (以前的) size. 6．The door opened and in came a troop of children in all sorts of fancy (奇特的) dresses. 7．Some animals are hunted illegally, so they become rare (珍贵的；稀有的)． 8．To make my room look nicer, I had the walls decorated (装饰) with paper cuts last year. Ⅱ.语境语法填空 1．?Having_removed (remove) from his hometown to the city, Zhang Qiang didn't know what ?to_do (do) at first.Under the guidance of his friend Li Ping, he made ?a fortune by doing business.He launched a campaign which ?was_designed (design) to help those in need in return ?for the society.He has done so much for the poor in his community that we can't think ?highly (high) of him enough. 2．As is known to all, cultural relics belong ?to human beings rather than individuals, so they are well worth ?protecting (protect)．However, some people are in search ?of them and take possession of them illegally.Some of them have been damaged while only a few survive.Therefore, every one of us should make all efforts to protect them from ?being_destroyed (destroy)．There is no doubt ?that it is our duty to protect cultural relics. 3．The old piano I bought years ago took up too much room, so ?by the light of the room, I ?took (take) it apart.I sold it to a waste recycling center at a low price. Ⅲ.语境改错 1．文中共有3处错误，每句中最多有两处，请找出并改正。 After removed dirt from the vase, the expert was amazed to find that it survived the Tang Dynasty.The vase was so amazingly designed that it was very worth buying at such a reasonable price. 答案：第一句：removed→removing; survived后加from 第二句：very→well 2．文中共有3处错误，每句中最多有两处，请找出并改正。 汽车英文演讲稿 篇一：汽车英语演讲稿 Good morning everyone ,today, the topic of my lecture（演讲） is “Do you love automobiles（'tmbilz）”. before I start my speech ,I want to ask a question,did there anybody present had seen the movie”the Fast and the Furious['fjrs]”.There are many cars of different styles in the movie,for example ,when you see the movie,you can find roadsters（跑车） just like Porsche（保时捷） GT3,sports utility (通用的）vehicles （车）SUV such as Volkswagen ['f:lks,va:gn]Touareg ['twɑ:reɡ](大众途锐）,even armored cars ,all of them are so powerful and beautiful. When I was a child,I dreamed I can have a car like FORD GT 40,this racing bike（跑车） is amazing and fantastic,but as time goes on ,my hobby is changing,now my favorite car is Jeep Rubicon,this is a car I really want ,it’s dynamic performance（性能） and off-road （越野） performance is top-ranking（一流的） ,which other ordinary SUV is unable to compare; the shape and color of this car is magnificent (华丽的） [mg'nfs()nt],and it’s gear-shift system（齿轮转 小学一年级介绍北京的英语作文 Beijing is an ancient city with a long history. Back in 3000 years ago in Zhou dynasty, Beijing, which was called Ji at the moment, had been named capital of Yan. Thereafter, Liao, Jin, Yuan, Ming and Qing dynasty all made Beijing their capital. Therefore, Beijing was famous for "Capital of a thousand years". The long history leaves Beijing precious cultural treasure. Winding for several kilometers in Beijing area, the Great Wall is the only man-made structure that could been seen in the space. The Summer Palace is a classic composition of ancient royal gardens, and the Forbidden City is the largest royal palaces in the world. Tiantan is where the emperor used to fete their ancestors, and also the soul of Chinese ancient constructions. The four sites above has been confirmed world cultural heritage by UNESCO. However, the best representatives for Beijing are the vanishing Hutongs and square courtyards. Through hundreds of years, they have become symbol of Beijing's life. Tian'anmen square being still brilliant today with cloverleaf junctions and skyscrapers everywhere, the old-timey scene and modern culture are combined to present a brand new visage of Beijing. As Beijing has been confirmed home city of Olympics 2008, the spirit of "green Olympics, scientific Olympics and humanized Olympic" will surely bring more and more changes to Beijing, promote the development of sports and Olympics in China as well as in the world, and strengthen the friendly communications between Chinese and foreign people. 北京是一个有着悠久历史的古城。 早在 3000 年前的周朝,北京,这叫霁,被命 名为首都燕。此后,辽、金、元、明、清都是北京首都。因此,北京著名的一千年 的“资本”。 悠久的历史使北京宝贵的文化瑰宝。绕组在北京地区几公里,长城是唯一的 人造结构,可以在空间。 颐和园是古代皇家园林的经典组合,和故宫是世界上最大 的皇家宫殿。 天坛是皇帝用来祭祀他们的祖先的地方,也是中国古代建筑的灵魂。 上面的四个网站已经确认被联合国教科文组织世界文化遗产。然而,北京最好的 代表是消失的胡同和广场庭院。数百年来,他们已经成为北京的生活的象征。天 安门广场到处都在今天依然灿烂的蝶式路口和摩天大楼的,古色古香的场景和现 代文化相结合,提出一个全新的北京的面貌。 随着北京 2008 年奥运会已被证实的家乡,精神的“绿色奥运、科技奥介绍北京的英语作文1篇 .doc
人教版二年级下册语文每课作业完整版
统编版三年级语文下册20.肥皂泡课后练习题作业(有答案)
汽车的重要性《英语作文》
我想去北京英语作文
必修2 Unit 1 课下作业(一~三)
汽车英文演讲稿
小学一年级介绍北京的英语作文
1/5