A5006519552- Multimodal Machine Learning Applications
- Advanced Image and Video Retrieval Techniques
- Handwritten Text Recognition Techniques
- Topic Modeling
- Pesticide Residue Analysis and Safety
- Domain Adaptation and Few-Shot Learning
- Speech and dialogue systems
- Multimedia Communication and Technology
- Pharmacological Effects and Assays
- Analytical chemistry methods development
- Digital Media Forensic Detection
- Machine Learning and Algorithms
- Video Analysis and Summarization
- Advanced Neural Network Applications
- Visual Attention and Saliency Detection
- Nuclear Physics and Applications
- Augmented Reality Applications
- Interactive and Immersive Displays
- Image and Object Detection Techniques
- Agricultural pest management studies
- Insect Resistance and Genetics
- Speech and Audio Processing
- Persona Design and Applications
- Image Processing and 3D Reconstruction
- Virtual Reality Applications and Impacts
Indian Agricultural Research Institute
2023-2024
Adobe Systems (United States)
2017-2023
Samsung (India)
2021
We introduce LEAF-QA, a comprehensive dataset of 250 <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">,</sub> 000 densely annotated figures/charts, constructed from real-world open data sources, along with 2 million question-answer (QA) pairs querying the structure and semantics these charts. LEAF-QA highlights problem multimodal QA, which is notably different conventional visual QA (VQA), has recently gained interest in community. Furthermore,...
This work summarizes the results of first Competition on Harvesting Raw Tables from Infographics (ICDAR 2019 CHART-Infographics). The complex process automatic chart recognition is divided into multiple tasks for purpose this competition, including Chart Image Classification (Task 1), Text Detection and Recognition 2), Role 3), Axis Analysis 4), Legend 5), Plot Element 6.a), Data Extraction 6.b), End-to-End 7). We provided a large synthetic training set evaluated submitted systems using...
Chart Question Answering (CQA) is the task of answering natural language questions about visualisations in chart image. Recent solutions, inspired by VQA approaches, rely on image-based attention for question/answering while ignoring inherent structure. We propose STL-CQA which improves through sequential elements localization, question encoding and then, a structural transformer-based learning approach. conduct extensive experiments proposing pre-training tasks, methodology also an improved...
We present a novel method, SALAD, for the challenging vision task of adapting pre-trained "source" domain network to "target" domain, with small budget annotation in and shift label space. Further, assumes that source data is not available adaptation, due privacy concerns or otherwise. postulate such systems need jointly optimize dual (i) selecting fixed number samples from target (ii) transfer knowledge domain. To do this, SALAD consists Guided Attention Transfer Network (GATN) an active...
With the explosion of video content on Internet, there is a need for research methods analysis which take human cognition into account. One such cognitive measure memorability, or ability to recall visual after watching it. Prior has looked image memorability and shown that it intrinsic content, but problem modeling not been addressed sufficiently. In this work, we develop prediction model including complexities in Detailed feature reveals proposed method correlates well with existing...
Online TV has seen rapid growth in recent years, with most of the large media companies broadcasting their linear content online. Access to online accounts is protected by an authentication, and like traditional cable subscription, users same household share credentials. However, as standard data collection techniques have capability collect only account level information, measurements fail capture individual viewing characteristics shared accounts. Thus, profile identification experience...
Voice assistants (VA) are finding a place in many households, with increasing numbers. Nevertheless, for every interaction user invokes VA using key or wake-up word, which is too common and hinders natural conversation. To solve this, we propose an On-Device solution that listens to the continuously only predefined period classifies utterance into device-directed non-device-directed deep learning-based model. Since our On-Device, privacy of maintained. We tried false acceptance as command...
<title>Abstract</title> Fall armyworm (FAW), <italic>Spodoptera frugiperda</italic> (J.E. Smith), a threat to maize production systems, is highly polyphagous pest of global significance. As per the National robotics policy for application drones in agriculture India, comparative study residue dynamics between drone and conventional prepared premix [Chlorantraniliprole (Chl) Emamectin benzoate (EB)] liquid formulation (CEOD), at 70 g (T1) 140 (T2) /ha two stages rabi plant was carried out....
A robust method was developed using LC-ESI-MS/MS-based identification and quantification of 103 fortified pesticides in a mango fruit drink. Variations QuEChERS extraction (without buffer, citrate, and/or acetate buffered) coupled with dispersive clean-up combinations were evaluated. Results showed 5 mL dilution citrate buffered anhydrous (anhy) MgSO 4 gave acceptable recovery for 100 @ 1 μg −1 fortification. The validated as per SANTE guidelines (SANTE/11813/2021). 95, 91, 77 satisfactorily...
Augmented Reality (AR) is rapidly gaining popularity, enhancing human perception of the real world by augmenting digital experiences. Existing tools for authoring AR scenes are either template based or require domain knowledge from experts, and therefore restrictive. ARComposer a novel interface that enables easy experiences free-form text describing scene. Our proposed allows creators to compose varied comprising multiple objects with diverse relationships each other as well models...
Documents are central to many business systems, and include forms, reports, contracts, invoices or purchase orders. The information in documents is typically natural language, but can be organized various layouts formats. There have been recent spurt of interest understanding document content with novel deep learning architectures. However, tasks need dense annotations, which costly scale generalize. Several active techniques proposed reduce the overall budget annotation while maintaining...
A method was developed using liquid chromatography tandem mass spectroscopy for the identification and quantification of multi residues pesticides. The present study is first this kind, destined purely to understand interaction between clean-up agents with 103 QuEChERS clean-up, employing most commonly used like anhyd.MgSO4, PSA, C-18 GCB in twelve combinations, performed assess their adsorption behavior Recovery studies at 1μg∙mL−1 showed that anhyd.MgSO4 gave acceptable recovery 100...