Facial micro-expression (ME) recognition has posed a huge challenge to researchers for its subtlety in motion and limited databases. Recently, handcrafted techniques have achieved superior performance but at the cost of domain specificity cumbersome parametric tunings. In this paper, we propose an Enriched Long-term Recurrent Convolutional Network (ELRCN) that first encodes each frame into feature vector through CNN module(s), then predicts by passing Long Short-term Memory (LSTM) module....

Micro-expression recognition is still in the preliminary stage, owing much to numerous difficulties faced development of datasets. Since micro-expression an important affective clue for clinical diagnosis and deceit analysis, effort has gone into creation these datasets research purposes. There are currently two publicly available spontaneous datasets--SMIC CASME II, both with baseline results released using widely used dynamic texture descriptor LBP-TOP feature extraction. Although popular...

One-stage object detectors are trained by optimizing classification-loss and localization-loss simultaneously, with the former suffering much from extreme foreground-background class imbalance issue due to large number of anchors. This paper alleviates this proposing a novel framework replace classification task in one-stage ranking task, adopting Average-Precision loss (AP-loss) for problem. Due its non-differentiability non-convexity, AP-loss cannot be optimized directly. For purpose, we...

Automatic facial micro-expression (ME) analysis is a growing field of research that has gained much attention in the last five years. With many recent works testing on limited data, there need to spur better approaches are both robust and effective. This paper summarises 2nd Facial Micro-Expression Grand Challenge (MEGC 2019) held conjunction with 14th IEEE Conference Face Gesture Recognition (FG) 2019. In this workshop, we proposed challenges for two tasks- spotting recognition, aim...

Micro-expressions are spontaneous, brief and subtle facial muscle movements that exposes underlying emotions. Motivated by recent exploits into deep learning for micro-expression analysis, we propose a lightweight dual-stream shallow network in the form of pair truncated CNNs with heterogeneous input features. The merging convolutional features allows discriminative classes stemming from both streams. Using activation heatmaps, further demonstrate salient areas well emphasized, correspond...

Micro-expression usually occurs at high-stakes situations and may provide useful information in the field of behavioral psychology for better interpretion analysis. Unfortunately, it is technically challenging to detect recognize micro-expressions due its brief duration subtle facial distortions. Apex frame, which instant indicating most expressive emotional state a video, effective classify emotion that particular frame. In this work, we present novel method spot apex frame spontaneous...

One-stage object detectors are trained by optimizing classification-loss and localization-loss simultaneously, with the former suffering much from extreme foreground-background class imbalance issue due to large number of anchors. This paper alleviates this proposing a novel framework replace classification task in one-stage ranking task, adopting Average-Precision loss (AP-loss) for problem. Due its non-differentiability non-convexity, AP-loss cannot be optimized directly. For purpose, we...

Few-shot action recognition aims to recognize novel classes (query) using just a few samples (support). The majority of current approaches follow the metric learning paradigm, which learns compare similarity between videos. Recently, it has been observed that directly measuring this is not ideal since different instances may show distinctive temporal distribution, resulting in severe misalignment issues across query and support In paper, we arrest problem from two distinct aspects --...

Abstract This paper focuses on self-supervised video representation learning. Most existing approaches follow the contrastive learning pipeline to construct positive and negative pairs by sampling different clips. However, this formulation tends bias static background has difficulty establishing global temporal structures. The major reason is that pairs, i.e., clips sampled from same video, have limited receptive fields, usually share similar backgrounds but differ in motions. To address...

In the recent year, state-of-the-art for facial micro-expression recognition have been significantly advanced by deep neural networks. The robustness of learning has yielded promising performance beyond that traditional handcrafted approaches. Most works in literature emphasized on increasing depth networks and employing highly complex objective functions to learn more features. this paper, we design a Shallow Triple Stream Three-dimensional CNN (STSTNet) is computationally light whilst...

Spontaneous subtle emotions are expressed through micro-expressions, which tiny, sudden and short-lived dynamics of facial muscles; thus poses a great challenge for visual recognition. The abrupt but significant the recognition task temporally sparse while rest, irrelevant dynamics, redundant. In this work, we analyze enforce sparsity constrains to learn temporal spectral structures eliminate would ease in spontaneous emotions. hypothesis is confirmed experimental results automatic emotion...

This paper addresses neural network based post-processing for the state-of-the-art video coding standard, High Efficiency Video Coding (HEVC). We first propose a partition-aware Convolution Neural Network (CNN) that utilizes partition information produced by encoder to assist in post-processing. In contrast existing CNN-based approaches, which only take decoded frame as input, proposed approach considers unit (CU) size and combines it with distorted such artifacts introduced HEVC are...

Body shape is about proportion, and fashion style all dressing those proportions to look their very best. Figuring out the styles suit a body can be daunting task for many people. It is, therefore, essential develop framework learning compatibility of shapes clothing styles. Though designers stylists have analyzed correlation between human long time, this issue did not receive much attention in multimedia science. In paper, we present novel recommender, on basis user's attributes. The rich...

Human-Object Interaction (HOI) detection recognizes how persons interact with objects, which is advantageous in autonomous systems such as self-driving vehicles and collaborative robots. However, current HOI detectors are often plagued by model inefficiency unreliability when making a prediction, consequently limits its potential for real-world scenarios. In this paper, we address these challenges proposing ERNet, an end-to-end trainable convolutional-transformer network detection. The...