Abstract. In this paper we propose a transparent, computationally simple and fast spatial domain ... more Abstract. In this paper we propose a transparent, computationally simple and fast spatial domain data hiding technique for content (image and audio) authentication purposes. It can also be used in data captioning, copyright protection, confidentiality and integrity over the property ...
International Conference on Visual Information Engineering (VIE 2003). Ideas, Applications, Experience, 2003
... 4. Figure 4 Provision of hasic augmented information for sport events. OBJECT TRACKING ALGORI... more ... 4. Figure 4 Provision of hasic augmented information for sport events. OBJECT TRACKING ALGORITHM The ... analysis. ACKNOWLEDGMENTS This research is supported by the European Union project MELISA. REFERENCES I ...
In this paper, a fast and memory-efficient listless version of set-partitioned embedded block (SP... more In this paper, a fast and memory-efficient listless version of set-partitioned embedded block (SPECK) image coder is proposed. Due to the use of linked lists, the original SPECK algorithm requires large run-time memory, making it unsuitable for memory-constrained applications. The proposed coder replaces linked list with small fixed size static memory, to keep track of blocks in set partitioning only and use only two markers to facilitate coding. Replacement of linked lists with small fixed-size static memory also reduces the memory access time, thereby making it faster than the original SPECK. Simulation results show that the proposed algorithm outperforms SPECK image coder in terms of memory requirement and computational complexity while retaining the coding efficiency and scalability. The proposed coder requires 0.25-bit per pixel state memory which is 3.125% of the memory to store the image. These features make it suitable for resource-constrained portable handheld multimedia de...
TENCON 2019 - 2019 IEEE Region 10 Conference (TENCON)
In this paper, a low complexity compressive sensing (CS) based scalable image codec is proposed. ... more In this paper, a low complexity compressive sensing (CS) based scalable image codec is proposed. At the encoder side, hierarchical wavelet decomposition of an input image is used that results in one low-frequency subband with no sparsity and a number of high-frequency subbands having sparsity. An adaptive measurement process is then applied on detail subbands to select the number of measurements. Set Partitioning in Hierarchical Tree (SPIHT) is used to encode the measurements to get a scalable bitstream. At the decoder, a non-iterative pseudoinverse algorithm with low complexity for CS reconstruction is used. Results show that the proposed method gives a PSNR improvement of 1 to 3.5 dB compared to other algorithms. Moreover, being non-iterative, the proposed codec is also having extremely low complexity.
2021 International Conference on Computing, Communication, and Intelligent Systems (ICCCIS), 2021
Noise is a commonly encountered distortion which generally affects the high frequency regions in ... more Noise is a commonly encountered distortion which generally affects the high frequency regions in images. It can lead to masking effect when the level of noise is high. To distinguish between natural and noise-afflicted images and to quantify the level of degradation, several statistical features have proved to be noteworthy. In this work, a method is proposed utilizing divisive normalization based on Natural Scenes Statistics (NSS) model which closely relates to human visual perception. It extracts features from spatial as well as frequency domain. Extracted features are used to drive the Machine Learning (ML) model Gaussian Process Regression (GPR) for mapping of the scores. Several methods have been proposed till date which are compared on three databases LIVE, CSIQ and TID2013. The trained model gives the highest correlation to Human Visual System for assessment of noise in natural images.
2017 4th IEEE Uttar Pradesh Section International Conference on Electrical, Computer and Electronics (UPCON), 2017
The Discrete Wavelet Transform (DWT) is extensively used for image coding due to its excellent en... more The Discrete Wavelet Transform (DWT) is extensively used for image coding due to its excellent energy compaction property and its ability to simultaneously analyze images in space-frequency domains. However, conventional methods of computing the DWT coefficients of an image require large amounts of memory, thus making them unsuitable for memory-constraint low-cost portable devices. In this paper we propose a novel low memory approach named Segmented Fractional Wavelet Filter SFrWF to compute the DWT of high resolution images on low-memory devices. Evaluation results show that the SFrWF requires less than 10 kB of RAM for a gray-scale image of size 2048×2048 thus making the SFrWF suitable for low-cost visual sensor nodes.
TENCON 2019 - 2019 IEEE Region 10 Conference (TENCON), 2019
The segmented fractional wavelet filter (SFrWF) allows the use of the fractional wavelet filter (... more The segmented fractional wavelet filter (SFrWF) allows the use of the fractional wavelet filter (FrWF) on line segments, thus can easily be implemented on low-memory devices and sensor nodes to compute the discrete wavelet transform of images. However, the complexity of SFrWF is relatively high due to the use of overlap and add method used in it for avoiding the boundary discontinuities at the segment boundaries. The high complexity of SFrWF makes it unsuitable for low-cost visual sensors used in wireless visual sensor networks (WVSNs)/Internet of things (IoT). In this paper, a lifting-based implementation of SFrWF with 9/7 filter-bank is presented with the aim to reduce it's computational complexity. The proposed lifting based implementation requires fewer computations than SFrWF. Furthermore, the proposed implementation uses an alternative approach to avoid the border discontinuities at segment boundaries. Evaluation results show that for high-resolution ($2048\times 2048$) images, the complexity of proposed implementation is approximately 40% less than that of SFrWF, without any additional memory requirement.
2018 5th IEEE Uttar Pradesh Section International Conference on Electrical, Electronics and Computer Engineering (UPCON), 2018
A typical image coder generally consists of a transform stage followed by quantization and coding... more A typical image coder generally consists of a transform stage followed by quantization and coding stages. The memory requirement of an image coder would be the maximum of both the stages and complexity would be the sum of both stages. Due to large memory requirements, most of the existing image coders are unsuitable for their implementation on memory-constrained-platforms especially for high-resolution images. In this paper, we propose a low memory approach, Block-based Fractional Wavelet Filter (BFrWF), to compute wavelet transform coefficients of high-resolution images. Furthermore, BFrWF can be combined with low memory wavelet-based image coding algorithms to design low-memory image codec. Evaluation results show that the BFrWF requires less than 10 kB of RAM (available over most of the low-cost sensor nodes) even for high-resolution (HR) images, thus making it suitable for visual sensor networks. Moreover, the proposed BFrWF implemented with 8 blocks has 25.64% less complexity t...
The quality of images get degraded due to distortion at any stage of capturing, transmission, com... more The quality of images get degraded due to distortion at any stage of capturing, transmission, compression, and reproduction etc. Therefore, to evaluate the quality, a large number of objective image quality assessment (IQA) algorithms/metrics were developed and many more are forthcoming. Each of these algorithms has its own merits and demerits in terms of complexity and accuracy. In this paper, performance of some of the recently full reference and no reference IQA algorithms against specific distortion due to JPEG and JPEG2000 compression and overall distortion on three popular public databases TID2008, LIVE and CSIQ is investigated. To measure and compare the performance, Spearman's rank order correlation coefficient (SROCC), which is reported in most of the literature, is used. It is found that almost all the metrics including PSNR give good performance for specific distortion due to JPEG and JPEG2000 compression. However, the performance degrades when the algorithms are appl...
Image segmentation refers to the separation of objects from the background, and has been one of t... more Image segmentation refers to the separation of objects from the background, and has been one of the most challenging aspects of digital image processing. Practically it is impossible to design a segmentation algorithm which has 100% accuracy, and therefore numerous segmentation techniques have been proposed in the literature, each with certain limitations. In this paper, a novel Falling-Ball algorithm is presented, which is a region-based segmentation algorithm, and an alternative to watershed transform (based on waterfall model). The proposed algorithm detects the catchment basins by assuming that a ball falling from hilly terrains will stop in a catchment basin. Once catchment basins are identified, the association of each pixel with one of the catchment basin is obtained using multicriterion fuzzy logic. Edges are constructed by dividing image into different catchment basins with the help of a membership function. Finally closed contour algorithm is applied to find closed regions...
This paper proposes a simple low memory architecture for computing discrete wavelet transform (DW... more This paper proposes a simple low memory architecture for computing discrete wavelet transform (DWT) of high-resolution (HR) images on low-cost memory-constrained sensor nodes used in visual sensor networks (VSN) or Internet of Multimedia Things (IoMT). The main feature of the proposed architecture is the novel data scanning technique that makes memory requirement independent of the image size. The proposed architecture needs only (30 S ) words of memory, where S is the number of parallel processing units and a critical path delay (CPD) equal to the delay of a multiplier ( T m ). Furthermore, a multiplierless version of this architecture is also proposed which reduces the CPD to T a < T m (where T a is the delay of an adder). In order to evaluate their effectiveness, the proposed architectures are coded in HDL and implemented on same FPGA board. Their performance is also compared with other state-of-the-art low memory DWT architectures. The experimental results show the superiority of the proposed architectures in terms of memory and CPD compared to existing DWT architectures. Moreover, the reduction in CPD to T a indicates that the operating frequency can be scaled up by several factors and can be chosen depending upon the application. Compared to one of the best state-of-the-art DWT architecture, proposed multiplierless architecture (with S = 4) needs 57.37% less LUT’s and 64.39% less flip-flops for HR image of dimension 2048 × 2048. Moreover, the proposed architecture needs no LUTRAM and DSP, whereas the existing architecture requires 3264 LUTRAM and 24 DSP’s. Thus the proposed multiplierless architecture is superior to the existing state-of-the-art architecture and is suitable for IoMT/VSNs.
The progressive nature of the JPEG2000 coded bitstream allows the reconstruction of images of dif... more The progressive nature of the JPEG2000 coded bitstream allows the reconstruction of images of different qualities from a single coded bitstream. This feature is utilized in this work to estimate the mean-squared-error (MSE) of reconstructed images without requiring the original image. It is based on the fact that if the MSE between the original image and a lower quality image is known, the MSE for higher quality images can be estimated from a quality scalable bitstream. The proposed method is highly accurate and is very simple as no complex statistical modeling is needed. Therefore, it is suitable to measure the fidelity of JPEG2000 decoded images at any desired quality in a real-time scenario.
2016 International Conference on Advances in Computing, Communications and Informatics (ICACCI), 2016
In view of increasing use of video transmission over wireless channels, a lot of research efforts... more In view of increasing use of video transmission over wireless channels, a lot of research efforts have been made to trade-off the error resiliency and transmission bandwidth requirements. This in turn has led to development of various unequal error protection (UEP) schemes. In this paper, the performance of optimized 2-layered and 3-layered UEP schemes using hierarchical 64 QAM (64-HQAM) for wavelet-transformed images coded with set partitioning in hierarchical trees (SPIHT) algorithm and transmitted over additive white Gaussian noise (AWGN) channel is investigated. A look-up table (LUT) based approach to select the optimal values of modulation parameters of 64-HQAM for different channel conditions is suggested so as to reduce the computational complexity. Simulation results show that 3-layered optimized 64-HQAM-based UEP outperforms corresponding non-optimized UEP. Also, 3-layered optimized UEP scheme using 64-HQAM performs slightly better as compared to optimized 2-layered UEP scheme using 64-HQAM.
Abstract. In this paper we propose a transparent, computationally simple and fast spatial domain ... more Abstract. In this paper we propose a transparent, computationally simple and fast spatial domain data hiding technique for content (image and audio) authentication purposes. It can also be used in data captioning, copyright protection, confidentiality and integrity over the property ...
International Conference on Visual Information Engineering (VIE 2003). Ideas, Applications, Experience, 2003
... 4. Figure 4 Provision of hasic augmented information for sport events. OBJECT TRACKING ALGORI... more ... 4. Figure 4 Provision of hasic augmented information for sport events. OBJECT TRACKING ALGORITHM The ... analysis. ACKNOWLEDGMENTS This research is supported by the European Union project MELISA. REFERENCES I ...
In this paper, a fast and memory-efficient listless version of set-partitioned embedded block (SP... more In this paper, a fast and memory-efficient listless version of set-partitioned embedded block (SPECK) image coder is proposed. Due to the use of linked lists, the original SPECK algorithm requires large run-time memory, making it unsuitable for memory-constrained applications. The proposed coder replaces linked list with small fixed size static memory, to keep track of blocks in set partitioning only and use only two markers to facilitate coding. Replacement of linked lists with small fixed-size static memory also reduces the memory access time, thereby making it faster than the original SPECK. Simulation results show that the proposed algorithm outperforms SPECK image coder in terms of memory requirement and computational complexity while retaining the coding efficiency and scalability. The proposed coder requires 0.25-bit per pixel state memory which is 3.125% of the memory to store the image. These features make it suitable for resource-constrained portable handheld multimedia de...
TENCON 2019 - 2019 IEEE Region 10 Conference (TENCON)
In this paper, a low complexity compressive sensing (CS) based scalable image codec is proposed. ... more In this paper, a low complexity compressive sensing (CS) based scalable image codec is proposed. At the encoder side, hierarchical wavelet decomposition of an input image is used that results in one low-frequency subband with no sparsity and a number of high-frequency subbands having sparsity. An adaptive measurement process is then applied on detail subbands to select the number of measurements. Set Partitioning in Hierarchical Tree (SPIHT) is used to encode the measurements to get a scalable bitstream. At the decoder, a non-iterative pseudoinverse algorithm with low complexity for CS reconstruction is used. Results show that the proposed method gives a PSNR improvement of 1 to 3.5 dB compared to other algorithms. Moreover, being non-iterative, the proposed codec is also having extremely low complexity.
2021 International Conference on Computing, Communication, and Intelligent Systems (ICCCIS), 2021
Noise is a commonly encountered distortion which generally affects the high frequency regions in ... more Noise is a commonly encountered distortion which generally affects the high frequency regions in images. It can lead to masking effect when the level of noise is high. To distinguish between natural and noise-afflicted images and to quantify the level of degradation, several statistical features have proved to be noteworthy. In this work, a method is proposed utilizing divisive normalization based on Natural Scenes Statistics (NSS) model which closely relates to human visual perception. It extracts features from spatial as well as frequency domain. Extracted features are used to drive the Machine Learning (ML) model Gaussian Process Regression (GPR) for mapping of the scores. Several methods have been proposed till date which are compared on three databases LIVE, CSIQ and TID2013. The trained model gives the highest correlation to Human Visual System for assessment of noise in natural images.
2017 4th IEEE Uttar Pradesh Section International Conference on Electrical, Computer and Electronics (UPCON), 2017
The Discrete Wavelet Transform (DWT) is extensively used for image coding due to its excellent en... more The Discrete Wavelet Transform (DWT) is extensively used for image coding due to its excellent energy compaction property and its ability to simultaneously analyze images in space-frequency domains. However, conventional methods of computing the DWT coefficients of an image require large amounts of memory, thus making them unsuitable for memory-constraint low-cost portable devices. In this paper we propose a novel low memory approach named Segmented Fractional Wavelet Filter SFrWF to compute the DWT of high resolution images on low-memory devices. Evaluation results show that the SFrWF requires less than 10 kB of RAM for a gray-scale image of size 2048×2048 thus making the SFrWF suitable for low-cost visual sensor nodes.
TENCON 2019 - 2019 IEEE Region 10 Conference (TENCON), 2019
The segmented fractional wavelet filter (SFrWF) allows the use of the fractional wavelet filter (... more The segmented fractional wavelet filter (SFrWF) allows the use of the fractional wavelet filter (FrWF) on line segments, thus can easily be implemented on low-memory devices and sensor nodes to compute the discrete wavelet transform of images. However, the complexity of SFrWF is relatively high due to the use of overlap and add method used in it for avoiding the boundary discontinuities at the segment boundaries. The high complexity of SFrWF makes it unsuitable for low-cost visual sensors used in wireless visual sensor networks (WVSNs)/Internet of things (IoT). In this paper, a lifting-based implementation of SFrWF with 9/7 filter-bank is presented with the aim to reduce it's computational complexity. The proposed lifting based implementation requires fewer computations than SFrWF. Furthermore, the proposed implementation uses an alternative approach to avoid the border discontinuities at segment boundaries. Evaluation results show that for high-resolution ($2048\times 2048$) images, the complexity of proposed implementation is approximately 40% less than that of SFrWF, without any additional memory requirement.
2018 5th IEEE Uttar Pradesh Section International Conference on Electrical, Electronics and Computer Engineering (UPCON), 2018
A typical image coder generally consists of a transform stage followed by quantization and coding... more A typical image coder generally consists of a transform stage followed by quantization and coding stages. The memory requirement of an image coder would be the maximum of both the stages and complexity would be the sum of both stages. Due to large memory requirements, most of the existing image coders are unsuitable for their implementation on memory-constrained-platforms especially for high-resolution images. In this paper, we propose a low memory approach, Block-based Fractional Wavelet Filter (BFrWF), to compute wavelet transform coefficients of high-resolution images. Furthermore, BFrWF can be combined with low memory wavelet-based image coding algorithms to design low-memory image codec. Evaluation results show that the BFrWF requires less than 10 kB of RAM (available over most of the low-cost sensor nodes) even for high-resolution (HR) images, thus making it suitable for visual sensor networks. Moreover, the proposed BFrWF implemented with 8 blocks has 25.64% less complexity t...
The quality of images get degraded due to distortion at any stage of capturing, transmission, com... more The quality of images get degraded due to distortion at any stage of capturing, transmission, compression, and reproduction etc. Therefore, to evaluate the quality, a large number of objective image quality assessment (IQA) algorithms/metrics were developed and many more are forthcoming. Each of these algorithms has its own merits and demerits in terms of complexity and accuracy. In this paper, performance of some of the recently full reference and no reference IQA algorithms against specific distortion due to JPEG and JPEG2000 compression and overall distortion on three popular public databases TID2008, LIVE and CSIQ is investigated. To measure and compare the performance, Spearman's rank order correlation coefficient (SROCC), which is reported in most of the literature, is used. It is found that almost all the metrics including PSNR give good performance for specific distortion due to JPEG and JPEG2000 compression. However, the performance degrades when the algorithms are appl...
Image segmentation refers to the separation of objects from the background, and has been one of t... more Image segmentation refers to the separation of objects from the background, and has been one of the most challenging aspects of digital image processing. Practically it is impossible to design a segmentation algorithm which has 100% accuracy, and therefore numerous segmentation techniques have been proposed in the literature, each with certain limitations. In this paper, a novel Falling-Ball algorithm is presented, which is a region-based segmentation algorithm, and an alternative to watershed transform (based on waterfall model). The proposed algorithm detects the catchment basins by assuming that a ball falling from hilly terrains will stop in a catchment basin. Once catchment basins are identified, the association of each pixel with one of the catchment basin is obtained using multicriterion fuzzy logic. Edges are constructed by dividing image into different catchment basins with the help of a membership function. Finally closed contour algorithm is applied to find closed regions...
This paper proposes a simple low memory architecture for computing discrete wavelet transform (DW... more This paper proposes a simple low memory architecture for computing discrete wavelet transform (DWT) of high-resolution (HR) images on low-cost memory-constrained sensor nodes used in visual sensor networks (VSN) or Internet of Multimedia Things (IoMT). The main feature of the proposed architecture is the novel data scanning technique that makes memory requirement independent of the image size. The proposed architecture needs only (30 S ) words of memory, where S is the number of parallel processing units and a critical path delay (CPD) equal to the delay of a multiplier ( T m ). Furthermore, a multiplierless version of this architecture is also proposed which reduces the CPD to T a < T m (where T a is the delay of an adder). In order to evaluate their effectiveness, the proposed architectures are coded in HDL and implemented on same FPGA board. Their performance is also compared with other state-of-the-art low memory DWT architectures. The experimental results show the superiority of the proposed architectures in terms of memory and CPD compared to existing DWT architectures. Moreover, the reduction in CPD to T a indicates that the operating frequency can be scaled up by several factors and can be chosen depending upon the application. Compared to one of the best state-of-the-art DWT architecture, proposed multiplierless architecture (with S = 4) needs 57.37% less LUT’s and 64.39% less flip-flops for HR image of dimension 2048 × 2048. Moreover, the proposed architecture needs no LUTRAM and DSP, whereas the existing architecture requires 3264 LUTRAM and 24 DSP’s. Thus the proposed multiplierless architecture is superior to the existing state-of-the-art architecture and is suitable for IoMT/VSNs.
The progressive nature of the JPEG2000 coded bitstream allows the reconstruction of images of dif... more The progressive nature of the JPEG2000 coded bitstream allows the reconstruction of images of different qualities from a single coded bitstream. This feature is utilized in this work to estimate the mean-squared-error (MSE) of reconstructed images without requiring the original image. It is based on the fact that if the MSE between the original image and a lower quality image is known, the MSE for higher quality images can be estimated from a quality scalable bitstream. The proposed method is highly accurate and is very simple as no complex statistical modeling is needed. Therefore, it is suitable to measure the fidelity of JPEG2000 decoded images at any desired quality in a real-time scenario.
2016 International Conference on Advances in Computing, Communications and Informatics (ICACCI), 2016
In view of increasing use of video transmission over wireless channels, a lot of research efforts... more In view of increasing use of video transmission over wireless channels, a lot of research efforts have been made to trade-off the error resiliency and transmission bandwidth requirements. This in turn has led to development of various unequal error protection (UEP) schemes. In this paper, the performance of optimized 2-layered and 3-layered UEP schemes using hierarchical 64 QAM (64-HQAM) for wavelet-transformed images coded with set partitioning in hierarchical trees (SPIHT) algorithm and transmitted over additive white Gaussian noise (AWGN) channel is investigated. A look-up table (LUT) based approach to select the optimal values of modulation parameters of 64-HQAM for different channel conditions is suggested so as to reduce the computational complexity. Simulation results show that 3-layered optimized 64-HQAM-based UEP outperforms corresponding non-optimized UEP. Also, 3-layered optimized UEP scheme using 64-HQAM performs slightly better as compared to optimized 2-layered UEP scheme using 64-HQAM.
Uploads