References & Citations
Computer Science > Computer Vision and Pattern Recognition
Title: Evaluation Of Hidden Markov Models Using Deep CNN Features In Isolated Sign Recognition
(Submitted on 19 Jun 2020 (v1), last revised 10 May 2021 (this version, v2))
Abstract: Isolated sign recognition from video streams is a challenging problem due to the multi-modal nature of the signs, where both local and global hand features and face gestures needs to be attended simultaneously. This problem has recently been studied widely using deep Convolutional Neural Network (CNN) based features and Long Short-Term Memory (LSTM) based deep sequence models. However, the current literature is lack of providing empirical analysis using Hidden Markov Models (HMMs) with deep features. In this study, we provide a framework that is composed of three modules to solve isolated sign recognition problem using different sequence models. The dimensions of deep features are usually too large to work with HMM models. To solve this problem, we propose two alternative CNN based architectures as the second module in our framework, to reduce deep feature dimensions effectively. After extensive experiments, we show that using pretrained Resnet50 features and one of our CNN based dimension reduction models, HMMs can classify isolated signs with 90.15% accuracy in Montalbano dataset using RGB and Skeletal data. This performance is comparable with the current LSTM based models. HMMs have fewer parameters and can be trained and run on commodity computers fast, without requiring GPUs. Therefore, our analysis with deep features show that HMMs could also be utilized as well as deep sequence models in challenging isolated sign recognition problem.
Submission history
From: Hacer Yalim Keles [view email][v1] Fri, 19 Jun 2020 15:18:03 GMT (1627kb,D)
[v2] Mon, 10 May 2021 13:24:33 GMT (1627kb,D)
Link back to: arXiv, form interface, contact.