Semantic segmentation plays a vital role in computer vision tasks, enabling precise pixel-level understanding of images. In this paper, we present a comprehensive library for semantic segmentation, which contains implementations of popular segmentation models like SegNet, FCN, UNet, and PSPNet. We also evaluate and compare these models on several datasets, offering researchers and practitioners a powerful toolset for tackling diverse segmentation challenges.
The aim of this paper is to demonstrate that a state of the art feature matcher (LoFTR) can be made more robust to rotations by simply replacing the backbone CNN with a steerable CNN which is equivariant to translations and image rotations. It is experimentally shown that this boost is obtained without reducing performance on ordinary illumination and viewpoint matching sequences.
In this report, we present a theoretical support of the continual learning method \textbf{Elastic Weight Consolidation}, introduced in paper titled `Overcoming catastrophic forgetting in neural networks'. Being one of the most cited paper in regularized methods for continual learning, this report disentangles the underlying concept of the proposed objective function. We assume that the reader is aware of the basic terminologies of continual learning.
Deep Learning (DL) combined with advanced model optimization methods such as RC-NN and Edge2Train has enabled offline execution of large networks on the IoT devices. In this paper, we compare the modern Deep Learning (DL) based video analytics approaches with the standard Computer Vision (CV) based approaches and finally, discuss the best-suited approach for video analytics on IoT devices.
In this project, we leverage a trained single-letter classifier to predict the written word from a continuously written word sequence, by designing a word reconstruction pipeline consisting of a dynamic-programming algorithm and an auto-correction model. We conduct experiments to optimize models in this pipeline, then employ domain adaptation to explore using this pipeline on unseen data distributions.
3D object and hand pose estimation have huge potentials for Augmented Reality, to enable tangible interfaces, natural interfaces, and blurring the boundaries between the real and virtual worlds. In this chapter, we present the recent developments for 3D object and hand pose estimation using cameras, and discuss their abilities and limitations and the possible future development of the field.
Part of a series on the development of Early Career Researchers in the lab. The idea for the CV clinic came from the lab themselves. We had previously had a session on creating a research profile and a large part of that session was spent looking at CVs.
In this work, we try to implement Image Processing techniques in the area of autonomous vehicles, both indoor and outdoor. The challenges for both are different and the ways to tackle them vary too. We also showed deep learning makes things easier and precise. We also made base models for all the problems we tackle while building an autonomous car for Indian Institute of Space science and Technology.
This article describes how to deploy dihedral group theory to detect Facial Key Points (FKP) symmetry to recognize emotions. The method can be applied in many other areas which those have the same data texture.
A classification algorithm that combines the components of k-nearest neighbours and multilayer neural networks has been designed and tested. With this method the computational time required for training the dataset has been reduced substancially. Gabor filters were used for the feature extraction to ensure a better performance. This algorithm is tested with MNIST dataset and it will be integrated as a module in the object recognition software which is currently under development.
We show the surprising result that colors across a change in viewing condition (changing light color, shading and camera) are related by a homography. Our homography color correction application delivers improved color fidelity compared with the linear least-square.
This paper implements and compares different techniques for face detection and recognition. One is find where the face is located in the images that is face detection and second is face recognition that is identifying the person. We study three techniques in this paper: Face detection using self organizing map (SOM), Face recognition by projection and nearest neighbor and Face recognition using SVM.
In this paper, we present some results of evidential reasoning in understanding multispectral images of remote sensing systems. The Dempster-Shafer approach of combination of evidences is pursued to yield contextual classification results, which are compared with previous results of the Bayesian context free classification, contextual classifications of dynamic programming and stochastic relaxation approaches.
In this project, we first study the Gaussian-based hidden Markov random field (HMRF) model and its expectation-maximization (EM) algorithm. Then we generalize it to Gaussian mixture model-based hidden Markov random field. The algorithm is implemented in MATLAB. We also apply this algorithm to color image segmentation problems and 3D volume segmentation problems.
Nonlinear bilateral filters (BF) deliver a fine blend of computational simplicity and blur-free denoising. However, little is known about their nature, noise-suppressing properties, and optimal choices of filter parameters. Our study is meant to fill this gap-explaining the underlying mechanism of bilateral filtering and providing the methodology for optimal filter selection. Practical application to CT image denoising is discussed to illustrate our results.
The compound models of clutter statistics are found suitable to describe the nonstationary nature of radar backscattering from high-resolution observations. In this letter, we show that the properties of Mellin transform can be utilized to generate higher order moments of simple and compound models of clutter statistics in a compact manner.