Abhimanyu Bhowmik

Vision-based quantification of the mechanical properties of plant branches before manipulation can be valuable for automated horticulture tasks. Efficient manipulation cannot avoid collisions, and hard contact with plants can harm them or require overly cautious motion. Therefore, gaining awareness of a plant’s mechanical compliance will be an asset for many downstream robotic plant-handling and monitoring tasks, such as flower picking or selective harvesting of fruits and vegetables. Estimating properties such as stiffness is fundamentally challenging because it requires force measurements, which are unreliable, expensive, and time-consuming to collect at scale. We propose using artificial wind as excitation to estimate relative compliance, which captures how much each branch segment bends compared to its neighbors. We propose a vision-based pipeline to track points on the branch, smooth their trajectories, and estimate per-segment compliance from the resulting bending signals. We estimate the uncertainty of the compliance metric by using a moving-block bootstrap, which quantifies the stability of the result under temporal resampling. This proof of concept is demonstrated using two wind-excited plant videos of 1104 and 590 frames, respectively. The proposed pipeline recovers a proximal-to-distal relative motion gradient and an interpretable relative compliance map. Our code and second experiment are reported in the supplementary material, which can be found at https://abhimanyubhowmik.github.io/rce/.

Vision-based underwater robots can be useful in inspecting and exploring confined spaces where traditional sensors and preplanned paths cannot be followed. Sensor noise and situational change can cause significant uncertainty in environmental representation. Thus, this paper explores how to represent mapping inconsistency in vision-based sensing and incorporate depth estimation confidence into the mapping framework. The scene depth and the confidence are estimated using the RAFT-Stereo model and are integrated into a voxel-based mapping framework, Voxblox. Improvements in the existing Voxblox weight calculation and update mechanism are also proposed. Finally, a qualitative analysis of the proposed method is performed in a confined pool and in a pier in the Trondheim fjord. Experiments using an underwater robot demonstrated the change in uncertainty in the visualization.

There is a very significant knowledge gap between Artificial Intelligence (AI) and a multitude of industries that exist in today’s modern world. This is primarily attributable to the limited availability of resources and technical expertise. However, a major obstacle is that AI needs to be flexible enough to work in many different applications, utilising a wide variety of datasets through cloud computing. As a result, we developed a benchmark toolkit called CloudAISim to make use of the power of AI and cloud computing in order to satisfy the requirements of modern applications. The goal of this study is to come up with a strategy for building a bridge so that AI can be utilised in order to assist those who are not very knowledgeable about technological advancements. In addition, we modelled a healthcare application as a case study in order to verify the scientific reliability of the CloudAISim toolkit and simulated it in a cloud computing environment using Google Cloud Functions to increase its real-time efficiency. A non-expert-friendly interface built with an interactive web app has also been developed. Any user without any technical knowledge can operate the entire model, which has a 98% accuracy rate. The proposed use case is designed to put AI to work in the healthcare industry, but CloudAISim would be useful and adaptable for other applications in the future.

The majority of financial transactions are now conducted virtually around the world. The widespread use of credit cards and online transactions encourages fraudulent activity. Thus, one of the most demanding real-world challenges is fraud detection. Unbalanced datasets, in which there are a disproportionately high number of non-fraud samples compared to incidents of fraud, are one of the key obstacles to effective fraud detection. A further factor complicating the learning process for cutting-edge machine learning classifiers is how quickly fraud behaviour changes. Thus, in this study, we suggest an efficient fraud detection methodology. We propose a unique nonlinear embedded clustering to resolve imbalances in the dataset, followed by a Deep Belief Network for detecting fraudulent transactions. The proposed model achieved an accuracy of 94% with a 70:30 ratio of training-validation dataset.