Research
My interests broadly lie in the fields of computer vision, machine learning and learning with limited labeled data (mainly active learning). I am currently working towards understanding how deep active learning can be made learnable and transferable.
Publications
-
Akshay L Chandra†, S.V. Desai†, C. Devaguptapu†, Vineeth N Balasubramanian
On Initial Pools for Deep Active Learning
Preregistration Workshop at NeurIPS 2020 (PMLR Proceedings)
† Equal Contribution
-
Akshay L Chandra, S.V. Desai, Vineeth N Balasubramanian, S. Ninomiya, Wei Guo
Active Learning with Point Supervision for Cost-Effective Panicle Detection in Cereal Crops
BioMed Central Plant Methods Journal (BMC), 2020 [Impact Factor: 4.5]
-
Akshay L Chandra†, S.V. Desai†, Wei Guo, S. Ninomiya, Vineeth N Balasubramanian
An Adaptive Supervision Framework for Active Learning in Object Detection
British Machine Vision Conference (BMVC), 2019
† Equal Contribution
-
Akshay L Chandra†, S.V. Desai†, M. Hirafuji, S. Ninomiya, V N Balasubramanian, W. Guo EasyRFP: An Easy to Use Edge Computing Toolkit for Real-Time Field Phenotyping
Extended Abstract at CVPPP & ECCV Academic Demonstrations, 2020
† Equal Contribution
Other Projects
- Deep Active Learning Toolkit for Image Classification
A PyTorch toolkit, May 2021
Abstract:
A PyTorch toolkit with 8 popular deep active learning query methods which include DBAL, VAAL, Coreset, Ensembler Variation-Ratio and more. Our toolkit currently supports CIFAR10/100, MNIST, SVHN, Tiny Imagenet and Long-Tails versions of CIFAR10/100.
-
Computer Vision with Deep Learning for Plant Phenotyping in Agriculture
A Survey Article Published at Advanced Computing & Communications (ACC) India, March 2020
Abstract:
In light of growing challenges in agriculture with ever growing food demand across the world, efficient crop management techniques are necessary to increase crop yield. Precision agriculture techniques allow the stakeholders to make effective and customized crop management decisions based on data gathered from monitoring crop environments. Plant phenotyping techniques play a major role in accurate crop monitoring. Advancements in deep learning have made previously difficult phenotyping tasks possible. This survey aims to introduce the reader to the state of the art research in deep plant phenotyping.
-
Image & Bounding Box Annotation Slicer
An Object Detection Data Transformer, April 2019
Abstract:
This easy-to-use library is a data transformer mainly useful in Object Detection tasks. It slices images and their bounding box annotations into smaller tiles, both into specific sizes and into any arbitrarynumber of equal parts. It can also resize them, both by specific sizes and by a resizing/scaling factor.
-
Mouse Cursor Control With Facial Movements
An HCI Application, October 2018
Abstract:
This Human-Computer Interaction application in Python will allow you to control your mouse cursor with your facial movements, works with just your regular webcam. Its hands-free, no wearable hardware or sensors needed.
-
Robust Morse Code Converter (Incomplete)
A Fun Deep Learning Application, August 2018
Abstract:
This 4-in-1 application can convert Morse Code signalled in 4 different ways in real time. Namely, flashlight toggles, eye winking, hand gestures and mouse clicks.
-
Selfie Filters Using Facial Landmarks
A Fun Facial Keypoints Application, May 2018
Abstract:
This deep learning application in Python can put various sunglasses on a detected face (I am calling them ‘Selfie Filters’) by detecting the Facial Landmarks (15 unique points). These landmarks mark important areas of the face - the eyes, corners of the mouth, the nose, etc.
-
Alphabet Recognition Through Gestures
A Gesture Recognition Application, April 2018
Abstract:
This deep learning application in python recognizes alphabet through gestures captured real time on a webcam. The user is allowed to write the alphabet on the screen using an object-of-interest (a water bottle cap in this case).