Mukul Sarkar

Bio: Mukul Sarkar is an academic researcher from Indian Institute of Technology Delhi. The author has contributed to research in topics: Image sensor & Pixel. The author has an hindex of 13, co-authored 79 publications receiving 580 citations. Previous affiliations of Mukul Sarkar include IMEC & Delft University of Technology.

Ultrasound Image Segmentation: A Deeply Supervised Network With Attention to Boundaries

Abstract: Objective: Segmentation of anatomical structures in ultrasound images requires vast radiological knowledge and experience. Moreover, the manual segmentation often results in subjective variations, therefore, an automatic segmentation is desirable. We aim to develop a fully convolutional neural network (FCNN) with attentional deep supervision for the automatic and accurate segmentation of the ultrasound images. Method: FCNN/CNNs are used to infer high-level context using low-level image features. In this paper, a sub-problem specific deep supervision of the FCNN is performed. The attention of fine resolution layers is steered to learn object boundary definitions using auxiliary losses, whereas coarse resolution layers are trained to discriminate object regions from the background. Furthermore, a customized scheme for downweighting the auxiliary losses and a trainable fusion layer are introduced. This produces an accurate segmentation and helps in dealing with the broken boundaries, usually found in the ultrasound images. Results: The proposed network is first tested for blood vessel segmentation in liver images. It results in $F1$ score, mean intersection over union, and dice index of 0.83, 0.83, and 0.79, respectively. The best values observed among the existing approaches are produced by U-net as 0.74, 0.81, and 0.75, respectively. The proposed network also results in dice index value of 0.91 in the lumen segmentation experiments on MICCAI 2011 IVUS challenge dataset, which is near to the provided reference value of 0.93. Furthermore, the improvements similar to vessel segmentation experiments are also observed in the experiment performed to segment lesions. Conclusion: Deep supervision of the network based on the input-output characteristics of the layers results in improvement in overall segmentation accuracy. Significance: Sub-problem specific deep supervision for ultrasound image segmentation is the main contribution of this paper. Currently the network is trained and tested for fixed size inputs. It requires image resizing and limits the performance in small size images.

Integrated Polarization Analyzing CMOS Image Sensor for Material Classification

Abstract: Material classification is an important application in computer vision. The inherent property of materials to partially polarize the reflected light can serve as a tool to classify them. In this paper, a real-time polarization sensing CMOS image sensor using a wire grid polarizer is proposed. The image sensor consist of an array of 128 × 128 pixels, occupies an area of 5 × 4 mm2 and it has been designed and fabricated in a 180-nm CMOS process. We show that this image sensor can be used to differentiate between metal and dielectric surfaces in real-time due to the different nature in partially polarizing the specular and diffuse reflection components of the reflected light. This is achieved by calculating the Fresnel reflection coefficients, the degree of polarization and the variations in the maximum and minimum transmitted intensities for varying specular angle of incidence. Differences in the physical parameters for various metal surfaces result in different surface reflection behavior, influencing the Fresnel reflection coefficients. It is also shown that the image sensor can differentiate among various metals by sensing the change in the polarization Fresnel ratio.

Biologically Inspired CMOS Image Sensor for Fast Motion and Polarization Detection

Abstract: A complementary-metal-oxide semiconductor (CMOS) image sensor replicating the perception of vision in insects is presented for machine vision applications. The sensor is equipped with in-pixel analog and digital memories that allow in-pixel binarization in real time. The binary output of the pixel tries to replicate the flickering effect of an insect's eye to detect the smallest possible motion based on the change in state of each pixel. The pixel level optical flow generation reduces the need for digital hardware and simplifies the process of motion detection. A built-in counter counts the changes in states for each row to estimate the direction of the motion. The designed image sensor can also sense polarization information in real time using a metallic wire grid micropolarizer. An extinction ratio of 7.7 is achieved. The 1-D binary optical flow is shown to vary with the polarization angle of the incoming light ray. The image sensor consists of an array of 128 × 128 pixels, occupies an area of 5 × 4 mm2 and it is designed and fabricated in a 180-nm CMOS process.

Ultrasound Image Enhancement Using Structure Oriented Adversarial Network

Abstract: In this letter, we aim to develop a deep adversarial despeckling approach to enhance the quality of ultrasound images. Most of the existing approaches target a complete removal of speckle, which produces oversmooth outputs and results in loss of structural details. In contrast, the proposed approach reduces the speckle extent without altering the structural and qualitative attributes of the ultrasound images. A despeckling residual neural network (DRNN) is trained with an adversarial loss imposed by a discriminator. The discriminator tries to differentiate between the despeckled images generated by the DRNN and the set of high-quality images. Further to prevent the developed network from oversmoothing, a structural loss term is used along with the adversarial loss. Experimental evaluations show that the proposed DRNN outperforms the state-of-the-art despeckling approaches in terms of the structural similarity index measure, peak signal to noise ratio, edge preservation index, and speckle region's signal to noise ratio.

Integrated polarization-analyzing CMOS image sensor for detecting incoming light ray direction

Abstract: A CMOS image sensor with an integrated wire grid polarizer to sense the polarization of light is presented. The chip consists of an array of 128 by 128 pixels, it occupies an area of 5×4 mm2 and it has been designed and fabricated in a CMOS 180nm process. The integrated grid polarizer is oriented in various directions to compute the Stokes parameters which can be used to determine the degree of polarization of the incoming light ray.

Computer vision : a modern approach = 计算机视觉 : 一种现代的方法

Abstract: From the Publisher: The accessible presentation of this book gives both a general view of the entire computer vision enterprise and also offers sufficient detail to be able to build useful applications. Users learn techniques that have proven to be useful by first-hand experience and a wide range of mathematical methods. A CD-ROM with every copy of the text contains source code for programming practice, color images, and illustrative movies. Comprehensive and up-to-date, this book includes essential topics that either reflect practical significance or are of theoretical importance. Topics are discussed in substantial and increasing depth. Application surveys describe numerous important application areas such as image based rendering and digital libraries. Many important algorithms broken down and illustrated in pseudo code. Appropriate for use by engineers as a comprehensive reference to the computer vision enterprise.

Underwater Optical Wireless Communication

Abstract: Underwater wireless information transfer is of great interest to the military, industry, and the scientific community, as it plays an important role in tactical surveillance, pollution monitoring, oil control and maintenance, offshore explorations, climate change monitoring, and oceanography research. In order to facilitate all these activities, there is an increase in the number of unmanned vehicles or devices deployed underwater, which require high bandwidth and high capacity for information transfer underwater. Although tremendous progress has been made in the field of acoustic communication underwater, however, it is limited by bandwidth. All this has led to the proliferation of underwater optical wireless communication (UOWC), as it provides higher data rates than the traditional acoustic communication systems with significantly lower power consumption and simpler computational complexities for short-range wireless links. UOWC has many potential applications ranging from deep oceans to coastal waters. However, the biggest challenge for underwater wireless communication originates from the fundamental characteristics of ocean or sea water; addressing these challenges requires a thorough understanding of complex physio-chemical biological systems. In this paper, the main focus is to understand the feasibility and the reliability of high data rate underwater optical links due to various propagation phenomena that impact the performance of the system. This paper provides an exhaustive overview of recent advances in UOWC. Channel characterization, modulation schemes, coding techniques, and various sources of noise which are specific to UOWC are discussed. This paper not only provides exhaustive research in underwater optical communication but also aims to provide the development of new ideas that would help in the growth of future underwater communication. A hybrid approach to an acousto-optic communication system is presented that complements the existing acoustic system, resulting in high data rates, low latency, and an energy-efficient system.

A New Frontier of Printed Electronics: Flexible Hybrid Electronics.

Abstract: The performance and integration density of silicon integrated circuits (ICs) have progressed at an unprecedented pace in the past 60 years. While silicon ICs thrive at low-power high-performance computing, creating flexible and large-area electronics using silicon remains a challenge. On the other hand, flexible and printed electronics use intrinsically flexible materials and printing techniques to manufacture compliant and large-area electronics. Nonetheless, flexible electronics are not as efficient as silicon ICs for computation and signal communication. Flexible hybrid electronics (FHE) leverages the strengths of these two dissimilar technologies. It uses flexible and printed electronics where flexibility and scalability are required, i.e., for sensing and actuating, and silicon ICs for computation and communication purposes. Combining flexible electronics and silicon ICs yields a very powerful and versatile technology with a vast range of applications. Here, the fundamental building blocks of an FHE system, printed sensors and circuits, thinned silicon ICs, printed antennas, printed energy harvesting and storage modules, and printed displays, are discussed. Emerging application areas of FHE in wearable health, structural health, industrial, environmental, and agricultural sensing are reviewed. Overall, the recent progress, fabrication, application, and challenges, and an outlook, related to FHE are presented.

A Review of the Pinned Photodiode for CCD and CMOS Image Sensors

Abstract: The pinned photodiode is the primary photodetector structure used in most CCD and CMOS image sensors. This paper reviews the development, physics, and technology of the pinned photodiode.

Photon Enhanced Thermionic Emission for Solar Concentrator Systems

Abstract: Solar-energy conversion usually takes one of two forms: the 'quantum' approach, which uses the large per-photon energy of solar radiation to excite electrons, as in photovoltaic cells, or the 'thermal' approach, which uses concentrated sunlight as a thermal-energy source to indirectly produce electricity using a heat engine. Here we present a new concept for solar electricity generation, photon-enhanced thermionic emission, which combines quantum and thermal mechanisms into a single physical process. The device is based on thermionic emission of photoexcited electrons from a semiconductor cathode at high temperature. Temperature-dependent photoemission-yield measurements from GaN show strong evidence for photon-enhanced thermionic emission, and calculated efficiencies for idealized devices can exceed the theoretical limits of single-junction photovoltaic cells. The proposed solar converter would operate at temperatures exceeding 200 degrees C, enabling its waste heat to be used to power a secondary thermal engine, boosting theoretical combined conversion efficiencies above 50%.