Adaptive downsampling to improve image compression at low bit rates

We can improve image compression by downsampling the image prior to encoding process and estimated the missing portion after decoding at low bit rates. This paper presents a new algorithm in which, based on the adaptive decision of appropriate downsampling ratios, quantization steps and interpolation we can achieve high quality of coding by consideration of local visual significance at low bit rates and it acts as a conventional JPEG compression at high bit rates. The full resolution image can be restored from the DCT coefficients of the downsampled pixels followed by interpolation scheme at the decoder. The experiments have demonstrated better PSNR improvement over the existing techniques before the critical bit rate. In addition, the adaptive mode decision not only makes the critical bit rate less image-independent, but also automates the switching coders in variable bit-rate applications, since the algorithm turns to the standard JPEG method whenever it is necessary at higher bit rates.

Although JPEG technique is considered as the most popular image compression standard, it behaves high visual degradation at low bit rates. In this paper, an efficient DCT–based image compression technique is proposed to achieve high Compression Ratio (CR) with high quality at both high and low bit rates. This technique uses switching between JPEG compression technique at high bit rates and a novel Adaptive Lossy Image Compression (ALIC) technique at low bit rates. ALIC is proposed to overcome the drawbacks of JPEG technique at low bitrates. The performance of the proposed technique is analysed at low and high bit rates on both grey and colour images. Performances of both JPEG and ALIC techniques are analysed and compared. The experimental results reveal that the proposed ALIC technique achieves better CR with acceptable SNR in comparison with JPEG technique. Also, the resultant CR of ALIC technique can be considerably increased with a slight decrease of its PSNR. This decrease in PSNR does not result in a noticeable visual degradation of the compressed image. On the other hand, increasing the CR of JPEG technique results in a noticeable visual degradation due to the appearance of blocking effect in the reconstructed image. Thus, it is greatly recommended to use ALIC technique in the applications that require high CR with stable PSNR. ALIC is a general purpose technique that can be applied, not only on images, but also on any data source which uses Huffman coding to achieve better CR. Therefore, it is suitable for compression of text, image and video.

This paper proposes a video coding scheme for low bit rates environments such as video streaming in internet or mobile networks. Since the annoying artifacts at low bit rates result mostly from the coarse quantization of residual signals, we try to reduce this problem by introducing the hybrid resolution scheme and allocating more bits to the transform coefficients. The intra frames are encoded in the original resolution for their importance, and the inter frames are downsampled and encoded by H.264 with finer quantization steps. The images of original resolution are reconstructed by using the decoded images of reduced resolution and previously reconstructed frames. Each block of original resolution is reconstructed by one of four modes, according to the combination of interpolation of the decoded block of reduced resolution, fetching of the previously reconstructed block of original resolution block, and use of the residual images of original resolution. And, the selection of mode is decided by the R-D optimization. The final bit stream consists of side information, bit stream of the H.264 for the reduced frames, and the residual signals of original resolution. Experimental results show that the proposed method yields better performance than the usual H.264/AVC up to some critical bit rates and the PSNR difference increases up to 3.5 dB at low bit rates. In addition, since the motion estimation is performed with the downsampled frames, the computational load for the overall encoding is much reduced. The proposed video coding system is especially efficient at low bit rates environment, thus it is suitable for video streaming in internet and mobile networks.

Multi-scale network (MsSG-CNN) for joint image and saliency map learning-based compression

To transmit video contents over limited bandwidth network, video bitstreams may need to reduce the bit rate by encoding with coarse quantization parameters at the expense of degrading quality. At low bit rates, better coding quality can be achieved by downsampling the video prior to compression and upsampling later after decompression. In this paper, we present an adaptive downsampling/upsampling video coding scheme in order to achieve better video quality at low bit rates in terms of both measure and visual quality. In particular, appropriate downsampling directions/ratios and quantization step sizes are adaptively decided for encoding different regions of video frame with the consideration of local contents. Experimental results have shown the better performance of the proposed scheme over the regular coding and downsampling-based coding scheme with fixed downscaling ratio. In addition, the proposed scheme significantly raises the critical bit rate below which a downsampling-based coding scheme outperforms the regular coding.

A low bit rates image compression framework based on adaptive block downsampling and super resolution (SR) was presented. At the encoder side, the downsampling mode and quantization mode of each 16 × 16 macroblock are determined adaptively using the ratio distortion optimization method, then the downsampled macroblocks are compressed by the standard JPEG. At the decoder side, the sparse representation-based SR algorithm is applied to recover full resolution macroblocks from decoded blocks. The experimental results show that the proposed framework outperforms the standard JPEG and the state-of-the-art downsampling-based compression methods in terms of both subjective and objective comparisons. Specifically, the peak signal-to-noise ratio gain of the proposed framework over JPEG reaches up to 2 to 4 dB at low bit rates, and the critical bit rate to JPEG is raised to about 2.3 bits per pixel. Moreover, the proposed framework can be extended to other block-based compression schemes.

The discrete cosine transform (DCT) is well known for its highly efficient coding performance and is widely used in many image compression applications. However, in low bit rate coding, it produces undesirable block artifacts that are visually not pleasing. In addition, in many practical applications, faster computation and easier VLSI implementation of DCT coefficients are also important issues. The removal of the block artifacts and faster DCT computation are therefore of practical interest. In this paper, we investigate a modified DCT computation scheme, to be called the subband DCT (SB-DCT), that provides a simple, efficient solution to the reduction of the block artifacts while achieving faster computation. We have applied the new approach for the low bit rate coding and decoding of images. Simulation results on real images have verified the improved performance obtained using the proposed method over the standard JPEG method.

For low bit rate video compression, the quality of reconstructed video is usually poor. The high codec priority of region of interest (ROI) can improve image quality obviously. Nowadays, video segmentation methods are often used for extracting ROI, but these methods have high computational complexity and are not satisfied to real time communication. On the other hand, in most existing rate control algorithms, ROI can't select the low and high bit rate R-Q model adaptively. Aiming at these problems, in this paper, a simple and efficient approach of extracting ROI is proposed which can decrease the computational complexity of existing ROI extracting algorithms. Bits are distributed to ROI and non- ROI (NROI) respectively according to the image complexity and motion information. Moreover, the judgment criterion of distinguishing between low and high bit rate coding category is derived, which makes the algorithm select the R-Q model adaptively and decrease the rate control errors. In addition, the scheme of modifying the coding order of macro blocks (MBs) can enhance the objective image quality. Experiment results demonstrate that the proposed algorithm achieves a bit rate closer to the target, provides fewer skipped frames, and gets better objective and subjective image quality significantly compared with TMN7 and TMN8 algorithms.

Many RGB digital cameras use a Bayer color filter array. In these cameras a full-resolution image is obtained from a mosaicked image by interpolation, and then the full resolution image is encoded. Similarly, mosaicked color filter arrays and interpolation methods for multispectral images have been proposed in recent years. The resulting full-resolution mutispectral images need to be encoded efficiently. This paper presents a new image compression method for one-shot multispectral imaging system. The proposed approach encodes the mosaicked multi-spectral image before interpolation. The decoder interpolates the image after decoding to reconstruct a full-resolution image. The simulation results show that the proposed method outperforms a conventional method, which encodes the full-resolution image after interpolation, at almost all bit rates. We also propose a heuristic color filter array design for our method that leads to lower bit rate.

Medical ultrasound image compression using contextual vector quantization

The role of compression is vital in telemedicine for the storage and transmission of medical images. This work is based on Contextual Vector Quantization (CVQ) compression algorithm with codebook optimization by Simulated Annealing (SA) for the compression of CT images. The region of interest (foreground) and background are separated initially by region growing algorithm. The region of interest is encoded with low compression ratio and high bit rate; the background region is encoded with high compression ratio and low bit rate. The codebook generated from foreground and background is merged, optimized by simulated annealing algorithm. The performance of CVQ-SA algorithm was validated in terms of metrics like Peak to Signal Noise Ratio (PSNR), Mean Square Error (MSE) and Compression Ratio (CR), the result was superior when compared with classical VQ, CVQ, JPEG lossless and JPEG lossy algorithms. The algorithms are developed in Matlab 2010a and tested on real-time abdomen CT datasets. The quality of reconstructed image was also validated by metrics like Structural Content (SC), Normalized Absolute Error (NAE), Normalized Cross Correlation (NCC) and statistical analysis was performed by Mann Whitney U Test. The outcome of this work will be an aid in the field of telemedicine for the transfer of medical images.

An error probability analysis is performed for noncoherent detection of optical heterodyne frequency-shift keying (FSK) direct sequence spread-spectrum code-division multiple access (CDMA) signals corrupted by laser phase noise, receiver noise, and multiuser noise. Receiver noise is included in the analysis as an additive white Gaussian process. The multiuser noise is also modeled as a Gaussian process. An FSK-CDMA system with low, moderate, and high user bit rates as compared with the laser linewidth is analyzed over different spreading code lengths for increasing numbers of users. The low bit rate system performs poorly even when the spreading code is long and the number of simultaneous users is very small. On the other hand, a high bit rate system which uses a spreading code with 2/sup 11/ chips can support up to 75 users if the maximum allowed possibility of bit error is 10/sup -6/. >

Most audio compression formats are based on the idea of low bit rate transparent encoding. As these types of audio signals are starting to migrate from portable players with inexpensive headphones to higher quality home audio systems, it is becoming evident that higher bit rates may be required to maintain transparency. We propose a novel method that enhances low bit rate encoded audio segments by applying multiband audio resynthesis methods in a postprocessing stage. Our algorithm employs the highly flexible Generalized Gaussian mixture model which offers a more accurate representation of audio features than the Gaussian mixture model. A novel residual conversion technique is applied which proves to significantly improve the enhancement performance without excessive overhead. In addition, both cepstral and residual errors are dramatically decreased by a feature-alignment scheme that employs a sorting transformation. Some improvements regarding the quantization step are also described that enable us to further reduce the algorithm overhead. Signal enhancement examples are presented and the results show that the overhead size incurred by the algorithm is a fraction of the uncompressed signal size. Our results show that the resulting audio quality is comparable to that of a standard perceptual codec operating at approximately the same bit rate.

An embedded image coder that provides very effective progressive image transmission and a high degree of spatial scalability while achieving superior visual performance to SPIHT is presented. A new partitioning structure combined with scale-based wavelet weights achieves a significance reordering that codes more significant information belonging to the lower frequency bands earlier in the bitstream when compared to SPIHT. This reordering provides substantial improvements in progressive transmission, spatial scalability, and PSNR, especially at low bit rates. Further visual gains are achieved by exploiting human visual system (HVS) characteristics to weight wavelet coefficients according to their perceptual importance, causing a reordering of the coefficients. The activity selective SPIHT coder provides higher perceptual quality and higher PSNRs than SPIHT at low bit rates, and comparable visual quality with slightly lower PSNRs at high bit rates.

We show an embedded wavelet coder with a prioritization protocol, which provides an acceleration of benefit gain for quantizers of interest at extremely low bit rates. This means that decoded outputs provide better image quality at very low bit rates. By changing its risk attitude within a rational approach, which avoids certain forms of behavioral inconsistency, a quantifier may modify the gain in benefit that results from a particular bitstream candidate to be transmitted at a truncation time. At medium and high bit rates, quantizers exhibit only low risk tolerance, since they are aware that the next truncation time might be the last one. Since at extremely low bit rates the target bit rate may be far away, quantizers are able to exhibit higher risk tolerance, and as a consequence they will have greater possibilities of accelerating their benefit gain. The cooperation among subsets of the quantizers may be needed to increase the risk tolerance at very low bit rates within a rational approach and still prioritize first the more relevant pieces of information at each truncation time. The members of any coalition of quantizers can then negotiate a feasible change in the risk attitudes of the quantizers of the coalition that would benefit them all. The final risk tolerance of different quantizers comes from the balance of power among the coalitions of quantizers. The prioritization protocol chooses to transmit, at each truncation time, a bitstream for the quantizer that receives the highest payment (per coding bit) in a coalitional game, which minimizes the dissatisfaction of coalitions.