Image Complexity Guided Network Compression for Biomedical Image Segmentation

Compression is a standard procedure for making convolutional neural networks (CNNs) adhere to some specific computing resource constraints. However, searching for a compressed architecture typically involves a series of time-consuming training/validation experiments to determine a good compromise between network size and performance accuracy. To address this, we propose an image complexity-guided network compression technique for biomedical image segmentation. Given any resource constraints, our framework utilizes data complexity and network architecture to quickly estimate a compressed model which does not require network training. Specifically, we map the dataset complexity to the target network accuracy degradation caused by compression. Such mapping enables us to predict the final accuracy for different network sizes, based on the computed dataset complexity. Thus, one may choose a solution that meets both the network size and segmentation accuracy requirements. Finally, the mapping is used to determine the convolutional layer-wise multiplicative factor for generating a compressed network. We conduct experiments using 5 datasets, employing 3 commonly-used CNN architectures for biomedical image segmentation as representative networks. Our proposed framework is shown to be effective for generating compressed segmentation networks, retaining up to ≈95% of the full-sized network segmentation accuracy, and at the same time, utilizing ≈32x fewer network trainable weights (average reduction) of the full-sized networks.

A New Biomedical Image Denoising Method Using an Adaptive Multi-resolution Technique

In the world of digital image processing, image denoising plays a vital role, where the primary objective was to distinguish between a clean and a noisy image. However, it was not a simple task. As a consequence of everyone's understanding of the practical challenge, a variety of methods have been presented during the last few years. Of those, wavelet transformer-based approaches were the most common. But wavelet-based methods have their own limitations in image processing applications like shift sensitivity, poor directionality, and lack of phase information, and they also face difficulties in defining the threshold parameters. As a result, this study provides an image de-noising approach based on Bi-dimensional Empirical Mode Decomposition (BEMD). This project's main purpose is to disintegrate noisy images based on their frequency and construct a hybrid algorithm that uses existing de-noising techniques. This approach decomposes the noisy picture into numerous IMFs with residue, which were subsequently filtered independently based on their specific properties. To quantify the success of the proposed technique, a comprehensive analysis of the experimental results of the benchmark test images was conducted using several performance measurement matrices. The reconstructed image was found to be more accurate and pleasant to the eye, outperforming state-of-the-art denoising approaches in terms of PSNR, MSE, and SSIM.

HHO-Based Vector Quantization Technique for Biomedical Image Compression in Cloud Computing

In the present digital era, the exploitation of medical technologies and massive generation of medical data using different imaging modalities, adequate storage, management, and transmission of biomedical images necessitate image compression techniques. Vector quantization (VQ) is an effective image compression approach, and the widely employed VQ technique is Linde–Buzo–Gray (LBG), which generates local optimum codebooks for image compression. The codebook construction is treated as an optimization issue solved with utilization of metaheuristic optimization techniques. In this view, this paper designs an effective biomedical image compression technique in the cloud computing (CC) environment using Harris Hawks Optimization (HHO)-based LBG techniques. The HHO-LBG algorithm achieves a smooth transition among exploration as well as exploitation. To investigate the better performance of the HHO-LBG technique, an extensive set of simulations was carried out on benchmark biomedical images. The proposed HHO-LBG technique has accomplished promising results in terms of compression performance and reconstructed image quality.

An Oscillator without Linear Terms: Infinite Equilibria, Chaos, Realization, and Application

Oscillations and oscillators appear in various fields and find applications in numerous areas. We present an oscillator with infinite equilibria in this work. The oscillator includes only nonlinear elements (quadratic, absolute, and cubic ones). It is different from common oscillators, in which there are linear elements. Special features of the oscillator are suitable for secure applications. The oscillator’s dynamics have been discovered via simulations and an electronic circuit. Chaotic attractors, bifurcation diagrams, Lyapunov exponents, and the boosting feature are presented while measurements of the implemented oscillator are reported by using an oscilloscope. We introduce a random number generator using such an oscillator, which is applied in biomedical image encryption. Moreover, the security and performance analysis are considered to confirm the correctness of encryption and decryption processes.

Digital Image Analysis Tools Developed by the Indiana O’Brien Center

The scale and complexity of images collected in biological microscopy have grown enormously over the past 30 years. The development and commercialization of multiphoton microscopy has promoted a renaissance of intravital microscopy, providing a window into cell biology in vivo. New methods of optical sectioning and tissue clearing now enable biologists to characterize entire organs at subcellular resolution. New methods of multiplexed imaging support simultaneous localization of forty or more probes at a time. Exploiting these exciting new techniques has increasingly required biomedical researchers to master procedures of image analysis that were once the specialized province of imaging experts. A primary goal of the Indiana O’Brien Center has been to develop robust and accessible image analysis tools for biomedical researchers. Here we describe biomedical image analysis software developed by the Indiana O’Brien Center over the past 25 years.

A single latent channel is sufficient for biomedical image segmentation

Glottis segmentation is a crucial step to quantify endoscopic footage in laryngeal high-speed videoendoscopy. Recent advances in using deep neural networks for glottis segmentation allow a fully automatic workflow. However, exact knowledge of integral parts of these segmentation deep neural networks remains unknown. Here, we show using systematic ablations that a single latent channel as bottleneck layer is sufficient for glottal area segmentation. We further show that the latent space is an abstraction of the glottal area segmentation relying on three spatially defined pixel subtypes. We provide evidence that the latent space is highly correlated with the glottal area waveform, can be encoded with four bits, and decoded using lean decoders while maintaining a high reconstruction accuracy. Our findings suggest that glottis segmentation is a task that can be highly optimized to gain very efficient and clinical applicable deep neural networks. In future, we believe that online deep learning-assisted monitoring is a game changer in laryngeal examinations.