To address the increasing need for efficient and accurate content moderation, we propose an efficient and lightweight deep classification ensemble structure. Our approach is based on a combination of simple visual features, designed for high-accuracy classification of violent content with low false positives. Our ensemble architecture utilizes a set of lightweight models with narrowed-down color features, and we apply it to both images and videos. We evaluated our approach using a large dataset of explosion and blast contents and compared its performance to popular deep learning models such as ResNet-50. Our evaluation results demonstrate significant improvements in prediction accuracy, while benefiting from 7.64x faster inference and lower computation cost. While our approach is tailored to explosion detection, it can be applied to other similar content moderation and violence detection use cases as well. Based on our experiments, we propose a "think small, think many" philosophy in classification scenarios. We argue that transforming a single, large, monolithic deep model into a verification-based step model ensemble of multiple small, simple, and lightweight models with narrowed-down visual features can possibly lead to predictions with higher accuracy.
Computational experiments are exploited in finding a well-designed processing path to optimize material structures for desired properties. This requires understanding the interplay between the processing-(micro)structure-property linkages using a multi-scale approach that connects the macro-scale (process parameters) to meso (homogenized properties) and micro (crystallographic texture) scales. Due to the nature of the problem's multi-scale modeling setup, possible processing path choices could grow exponentially as the decision tree becomes deeper, and the traditional simulators' speed reaches a critical computational threshold. To lessen the computational burden for predicting microstructural evolution under given loading conditions, we develop a neural network (NN)-based method with physics-infused constraints. The NN aims to learn the evolution of microstructures under each elementary process. Our method is effective and robust in finding optimal processing paths. In this study, our NN-based method is applied to maximize the homogenized stiffness of a Copper microstructure, and it is found to be 686 times faster while achieving 0.053% error in the resulting homogenized stiffness compared to the traditional finite element simulator on a 10-process experiment.
The ability to distinguish between different movie scenes is critical for understanding the storyline of a movie. However, accurately detecting movie scenes is often challenging as it requires the ability to reason over very long movie segments. This is in contrast to most existing video recognition models, which are typically designed for short-range video analysis. This work proposes a State-Space Transformer model that can efficiently capture dependencies in long movie videos for accurate movie scene detection. Our model, dubbed TranS4mer, is built using a novel S4A building block, which combines the strengths of structured state-space sequence (S4) and self-attention (A) layers. Given a sequence of frames divided into movie shots (uninterrupted periods where the camera position does not change), the S4A block first applies self-attention to capture short-range intra-shot dependencies. Afterward, the state-space operation in the S4A block is used to aggregate long-range inter-shot cues. The final TranS4mer model, which can be trained end-to-end, is obtained by stacking the S4A blocks one after the other multiple times. Our proposed TranS4mer outperforms all prior methods in three movie scene detection datasets, including MovieNet, BBC, and OVSD, while also being $2\times$ faster and requiring $3\times$ less GPU memory than standard Transformer models. We will release our code and models.
Understanding the impact of tumor biology on the composition of nearby cells often requires characterizing the impact of biologically distinct tumor regions. Biomarkers have been developed to label biologically distinct tumor regions, but challenges arise because of differences in the spatial extent and distribution of differentially labeled regions. In this work, we present a framework for systematically investigating the impact of distinct tumor regions on cells near the tumor borders, accounting their cross spatial distributions. We apply the framework to multiplex immunohistochemistry (mIHC) studies of pancreatic cancer and show its efficacy in demonstrating how biologically different tumor regions impact the immune response in the tumor microenvironment. Furthermore, we show that the proposed framework can be extended to largescale whole slide image analysis.
Virtual Reality has become a significant element of education throughout the years. To understand the quality and advantages of these techniques, it is important to understand how they were developed and evaluated. Since COVID-19, the education system has drastically changed a lot. It has shifted from being in a classroom with a whiteboard and projectors to having your own room in front of your laptop in a virtual meeting. In this respect, virtual reality in the laboratory or Virtual Laboratory is the main focus of this research, which is intended to comprehend the work done in quality education from a distance using VR. As per the findings of the study, adopting virtual reality in education can help students learn more effectively and also help them increase perspective, enthusiasm, and knowledge of complex notions by offering them an interactive experience in which they can engage and learn more effectively. This highlights the importance of a significant expansion of VR use in learning, the majority of which employ scientific comparison approaches to compare students who use VR to those who use the traditional method for learning.
The development of the Automatic License Plate Recognition (ALPR) system has received much attention for the English license plate. However, despite being the sixth largest population around the world, no significant progress can be tracked in the Bengali language countries or states for the ALPR system addressing their more alarming traffic management with inadequate road-safety measures. This paper reports a computationally efficient and reasonably accurate Automatic License Plate Recognition (ALPR) system for Bengali characters with a new end-to-end DNN model that we call Bengali License Plate Network(BLPnet). The cascaded architecture for detecting vehicle regions prior to vehicle license plate (VLP) in the model is proposed to eliminate false positives resulting in higher detection accuracy of VLP. Besides, a lower set of trainable parameters is considered for reducing the computational cost making the system faster and more compatible for a real-time application. With a Computational Neural Network (CNN)based new Bengali OCR engine and word-mapping process, the model is characters rotation invariant, and can readily extract, detect and output the complete license plate number of a vehicle. The model feeding with17 frames per second (fps) on real-time video footage can detect a vehicle with the Mean Squared Error (MSE) of 0.0152, and the mean license plate character recognition accuracy of 95%. While compared to the other models, an improvement of 5% and 20% were recorded for the BLPnetover the prominent YOLO-based ALPR model and the Tesseract model for the number-plate detection accuracy and time requirement, respectively.
Vision-based deep learning models can be promising for speech-and-hearing-impaired and secret communications. While such non-verbal communications are primarily investigated with hand-gestures and facial expressions, no research endeavour is tracked so far for the lips state (i.e., open/close)-based interpretation/translation system. In support of this development, this paper reports two new Convolutional Neural Network (CNN) models for lips state detection. Building upon two prominent lips landmark detectors, DLIB and MediaPipe, we simplify lips-state model with a set of six key landmarks, and use their distances for the lips state classification. Thereby, both the models are developed to count the opening and closing of lips and thus, they can classify a symbol with the total count. Varying frame-rates, lips-movements and face-angles are investigated to determine the effectiveness of the models. Our early experimental results demonstrate that the model with DLIB is relatively slower in terms of an average of 6 frames per second (FPS) and higher average detection accuracy of 95.25%. In contrast, the model with MediaPipe offers faster landmark detection capability with an average FPS of 20 and detection accuracy of 94.4%. Both models thus could effectively interpret the lips state for non-verbal semantics into a natural language.
To facilitate implementation of high-accuracy deep neural networks especially on resource-constrained devices, maintaining low computation requirements is crucial. Using very deep models for classification purposes not only decreases the neural network training speed and increases the inference time, but also need more data for higher prediction accuracy and to mitigate false positives. In this paper, we propose an efficient and lightweight deep classification ensemble structure based on a combination of simple color features, which is particularly designed for "high-accuracy" image classifications with low false positives. We designed, implemented, and evaluated our approach for explosion detection use-case applied to images and videos. Our evaluation results based on a large test test show considerable improvements on the prediction accuracy compared to the popular ResNet-50 model, while benefiting from 7.64x faster inference and lower computation cost. While we applied our approach to explosion detection, our approach is general and can be applied to other similar classification use cases as well. Given the insight gained from our experiments, we hence propose a "think small, think many" philosophy in classification scenarios: that transforming a single, large, monolithic deep model into a verification-based step model ensemble of multiple small, simple, lightweight models with narrowed-down color spaces can possibly lead to predictions with higher accuracy.
Training a multimodal network is challenging and it requires complex architectures to achieve reasonable performance. We show that one reason for this phenomena is the difference between the convergence rate of various modalities. We address this by pre-training modality-specific sub-networks in multimodal architectures independently before end-to-end training of the entire network. Furthermore, we show that the addition of an attention mechanism between sub-networks after pre-training helps identify the most important modality during ambiguous scenarios boosting the performance. We demonstrate that by performing these two tricks a simple network can achieve similar performance to a complicated architecture that is significantly more expensive to train on multiple tasks including sentiment analysis, emotion recognition, and speaker trait recognition.