Abstract

With the increased importance of autonomous navigation systems has come an increasing need to protect the safety of Vulnerable Road Users (VRUs) such as pedestrians. Predicting pedestrian intent is one such challenging task, where prior work predicts the binary cross/no-cross intention with a fusion of visual and motion features. However, there has been no effort so far to hedge such predictions with humanunderstandable reasons. We address this issue by introducing a novel problem setting of exploring the intuitive reasoning behind a pedestrian’s intent. In particular, we show that predicting the ‘WHY’ can be very useful in understanding the ‘WHAT’. To this end, we propose a novel, reasonenriched PIE++ dataset consisting of multi-label textual explanations/reasons for pedestrian intent. We also introduce a novel multi-task learning framework called MINDREAD, which leverages a cross-modal representation learning framework for predicting pedestrian intent as well as the reason behind the intent. Our comprehensive experiments show significant improvement of 5.6% and 7% in accuracy and F1-score for the task of intent prediction on the PIE++ dataset using MINDREAD. We also achieved a 4.4% improvement in accuracy on a commonly used JAAD dataset. Extensive evaluation using quantitative/qualitative metrics and user studies shows the effectiveness of our approach.

Abstract

Prior works have addressed the problem of driver intention prediction (DIP) by identifying maneuvers after their onset. On the other hand, early anticipation is equally important in scenarios that demand a preemptive response before a maneuver begins. However, there is no prior work aimed at addressing the problem of driver action anticipation before the onset of the maneuver, limiting the ability of the advanced driver assistance system (ADAS) for early maneuver anticipation. In this work, we introduce Anticipating Driving Maneuvers (ADM), a new task that enables driver action anticipation before the onset of the maneuver. To initiate research in ADM task, we curate Driving Action Anticipation Dataset, DAAD, that is multi-view: in- and out-cabin views in dense and heterogeneous scenarios, and multimodal: egocentric view and gaze information. The dataset captures sequences both before the initiation and during the execution of a maneuver. During dataset collection, we also ensure to capture wide diversity in traffic scenarios, weather and illumination, and driveway conditions. Next, we propose a strong baseline based on a transformer architecture to effectively model multiple views and modalities over longer video lengths. We benchmark the existing DIP methods on DAAD and related datasets. Finally, we perform an ablation study showing the effectiveness of multiple views and modalities in maneuver anticipation

Abstract

This paper presents the competition report on Reading Documents through Aria Glasses (ICDAR 2024 RDTAG) held at the 18th International Conference on Document Analysis and Recognition (ICDAR 2024). From a mixed reality perspective, understanding the text in the world is of paramount importance. However, all day long, always on, machine perception devices like Aria Glasses pose a unique primary challenge of lower resolution due to their power and sensor constraints. Moreover, diverse everyday scenes like variations in the lighting conditions and reading positions further complicate the reading tasks. To address this, we propose a new dataset and a challenge. Specifically, we propose three novel tasks: Isolated Word Recognition in Low Resolution (Task A), Prediction of Reading Order (Task B), and Page Level Recognition (Task C). We provide new training and test sets consisting of document images captured by Aria Glasses while reading diverse documents in English under various everyday scenarios. Our aim is to engage researchers with prior experience in English language OCR, and to establish benchmarks contributing to the academic literature in this field. A total of thirty-three different teams from around the world registered for this competition, and twelve teams submitted their results along with algorithm details. The winning team, SRCB, achieved a 97.23% Character Recognition Rate (CRR) and a 90.45% Word Recognition Rate (WRR) for Task A: Isolated Word Recognition in Low Resolution. Team Gang-of-N won Task B: Prediction of Reading Order with a BLEU score of 0.0939. Team SRCB also won Task C: Page Level Recognition and Reading with a 77.44% average Page Level Character Recognition Rate (PCRR) and a 50.55% average Page Level Word Recognition Rate (PWRR).

Abstract

This paper presents a competition report on the Recognition and Visual Question Answer on Handwritten Documents towards deeper understanding of handwritten multilingual documents (ICDAR 2024-HWD) held at the 18th International Conference on Document Analysis and Recognition (ICDAR 2024). Documents are in English or Indian Languages. Earlier editions related to recognition of Indian handwriting were held in conjunction with ICFHR 2022 and ICDAR 2023. A related DocVQA task was held in DAS 2020. This edition proposes three main tasks: Isolated Word Recognition (Task A), Page Level Recognition and Reading (Task B), and Visual Question Answers on Handwritten Documents (Task A). While Task A was already part of our previous competitions, we bring in new data as part of this edition. Task B and Task C are novel additions for this year. By attracting researchers with experience in printed and handwritten documents, we aim to establish benchmarks that significantly contribute to the academic literature in this field. A total of thirty-two teams around the world registered for this competition. Among them, only ten teams submitted their results along with algorithm details. The winning team, TSNUK, achieved an average 98.00% Character Recognition Rate (CRR) and 94.26% Word Recognition Rate (WRR) across four languages for Task A: Isolated Word Recognition. IndependentOCR excelled in Task B: Page Level Recognition and Reading, with 76.32% average Page Level Character Recognition Rate (PCRR) and 62.57% average Page Level Word Recognition Rate. The team PA_VCG won Task C: Visual Question Answering on Handwritten Documents with a 0.643 ANLS score.

Abstract

Knowledge Graph Completion (KGC) aims to predict missing entities or relations in knowledge graph but it becomes computationally expensive as KG scales. Existing research focuses on bilinear pooling-based factorization methods (LowFER, TuckER) to solve this problem. These approaches introduce too many trainable parameters which obstruct the deployment of these techniques in many real-world scenarios. In this paper, we introduce a novel parameter-efficient framework, KGRP which a) approximates bilinear pooling using Kernelized Random Projection matrix b) employs CNN for the better fusion of entities and relations to infer missing links. Our experimental results show that KGRP has 73% fewer parameters as compared to the state-of-theart approaches (LowFER, TuckER) for the knowledge graph completion task while retaining 88% performance for the best baseline. Furthermore, we also provide novel insights on the interpretability of relation embeddings. We also test the effectiveness of KGRP on a large-scale recruitment knowledge graph of 0.25 M entities.

Abstract

There are certain dynamics while being non-Markovian, do never exhibit information backflow. We show that if two such dynamical maps are considered in a scenario where the order of application of these two dynamical maps are not definite, the effective channel can manifest information backflow. In particular, we use quantum SWITCH to activate such a channel. In contrast, activation of those channels are not possible even if one uses many copies of such channels in series or in parallel action. We then investigate the dynamics behind the quantum SWITCH experiment and find out that after the action of quantum SWITCH both the CP (Complete Positive)- divisibility and P (Positive)- divisibility of the channel breaks down, along with the activation of information backflow. Our study elucidate the advantage of quantum SWITCH by investigating its dynamical behavior.

Abstract

The coronavirus disease (COVID-19) caused over 170 million illnesses and over 3 million deaths worldwide. Researchers from all over the world have been using a variety of machine-learning techniques to classify the DNA sequences of the SARS-CoV-2 virus. To classify SARS-CoV-2 (Covid-19) whole genome sequences with other viruses including Dengue, Ebola, Influenza, and also Coronaviruses that may infect humans and are members of the family of Coronaviridae like SARS-CoV, MERS-CoV, we have proposed a multi-dimensional frequency encoding scheme. In the proposed method each genome is converted into four frequency-encoded sequences for each nucleotide. Then each encoded sequence is partitioned into quartiles and fed to the CNN. The findings of the binary classification of SARS-CoV-2 with other viruses and multi-class classification of six different viruses using five well-known machine learning algorithms and the proposed approach using CNN were reported in this study. Furthermore, we applied machine learning methods for the multi-class classification of eight different SARS-CoV-2 variants. The proposed approach performed well among all machine-learning techniques for classifying genome sequences of viruses and SARS-CoV-2 variants with accuracy of 99% and 98% respectively. In addition, we have also compared the performance of our proposed approach with the existing methods from the literature.

Abstract

Multivariate drought indices including various hydrological processes into account can be more valuable under climatic and anthropogenic changes. Standardized Precipitation Evapotranspiration Index (SPEI) and Standardized precipitation Actual Evapotranspiration Index (SPAEI) are the drought indices used to estimate drought index, considering precipitation, and evapotranspiration (ET) into account. Many studies used empirical, machine learning, and process-based model estimates of ET for the calculation of drought indices of SPEI and SPAEI. However, the sensitivity of ET estimates on drought characteristics at the catchment scale is highly complex. The present study aimed to include and analyze the sensitivity of various approaches of empirical (Budyko, Penman–Monteith, Hargreaves, and Turc), modeled (SWAT), and remote sensing (MODIS) in the drought characterization using SPEI and SPAEI. The present methodology was tested on a dry-sub-humid river catchment of India, the Tunga-Bhadra River catchment for the period of 2000–2012. The performance of statistical indicators (Nash–Sutcliffe Efficiency and R2) for SPEI values by various empirical methods of Potential Evapotranspiration (PET) (i) Penman–Monteith (ii) Hargreaves against remote sensing PET were 0.93 and 0.95, respectively, which are high in comparison with SWAT simulated PET-based SPEI values, which shows NSE values of 0.85 against remote sensing PET-based SPEI.

Abstract

Event-based models focus on modelling of peak runoff from rainfall data. Conceptual models indicate simplified models that provide reasonably accurate answers despite their crude nature. Rainfall-runoff models are used to transform a rainfall event into a runoff event. This paper focuses on reviewing computational simulation of rainfall-runoff processes over a catchment. Lumped conceptual, event-based rainfall-runoff models have remained the dominant practice for design flood estimation in Australia for many years due to their simplicity, flexibility, and accuracy under certain conditions. Attempts to establish regionalization methods for prediction of design flood hydrographs in ungauged catchments have seen little success. Therefore, as well as reviewing key rainfall-runoff model components for design flood estimation with a special focus on event-based conceptual models, this paper covers the aspects of regionalization to promote their applications to ungauged catchments.

Abstract

Atmospheric methane (CH4) is a potent climate change agent responsible for a fraction of global 20 warming. The present study investigated the spatial and temporal variability of atmospheric column-averaged (X) CH4 (XCH4) concentrations using Greenhouse gases Observing SATellite (GOSAT) and TROPOspheric Monitoring Instrument onboard the Sentienl-5 Precursor (S5P/TROPOMI) data from 2009 to 2022 over the South Asia region. During the study period, the long-term trends in XCH4 increased from 1700 ppb to 1950 ppb with an annual growth rate of 8.76 ppb year-1 25 . Among all natural and anthropogenic sources of CH4, the rate of increase in XCH4 was higher over the Mundra thermal power station and Mundra ultra mega power plant at about 9.62 ppb year-1 , followed by the coal site at about 8.76 ppb year-1 (Korba). With a growth rate of 8.61 ppb year-1 , the Sundarbans natural wetland competes with coal sites, producing over 30 MT, indicating an equivalent anthropogenic source. For the 15 Indian Agroclimatic zones, 30 significant high emissions of CH4 were observed over the Middle Gangetic Plains (MGP), Trans Gangetic Plains (TGP), Upper Gangetic Plains (UGP), East Coast Plains & Hills (ECPH), Lower Gangetic Plains (LGP) and East Gangetic Plains (EGP). Further, the bottom-up anthropogenic CH4 emissions data are mapped against the XCH4 concentrations and found high correlation in the Indo Gangetic Plains (IGP) region, indicating the hotspots of anthropogenic CH4. The present study 35 highlighted the impact of natural and anthropogenic sources of XCH4 and quantified the spatiotemporal changes in XCH4 at each study site over the Indian region