Comparative Analysis of Pre-trained based CNN-RNN Deep Learning Models on Anomaly-5 Dataset for Action Recognition
Action recognition in videos is one of the essential, challenging and active area of research in the field of computer vision that adopted in various applications including automated surveillance systems, security systems and human computer interaction. In this paper, we present an in-depth compara...
Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Sukkur IBA University
2024-10-01
|
| Series: | Sukkur IBA Journal of Computing and Mathematical Sciences |
| Online Access: | https://journal.iba-suk.edu.pk:8089/sibajournals/index.php/sjcms/article/view/1444 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1825197131881775104 |
|---|---|
| author | Fayaz Ahmed Memon Umair Ali Khan Pardeep Kumar Imtiaz Ali Halepoto Farida Memon |
| author_facet | Fayaz Ahmed Memon Umair Ali Khan Pardeep Kumar Imtiaz Ali Halepoto Farida Memon |
| author_sort | Fayaz Ahmed Memon |
| collection | DOAJ |
| description |
Action recognition in videos is one of the essential, challenging and active area of research in the field of computer vision that adopted in various applications including automated surveillance systems, security systems and human computer interaction. In this paper, we present an in-depth comparative analysis of five CNN-RNN models based on pre-trained networks such as InceptionV3, VGG16, MobileNetV2, ResNet152V2 and InceptionResNetV2 with recurrent LSTM units for action recognition on Anomaly-5 dataset. The performance of these models is analyzed and compared in terms of accuracy, precision, recall & F1-scores and computational efficiency. The CNN-RNN architectures we considered for analysis in this paper, the ResNet152V2 based CNN-RNN model exhibits better performance and achieved highest accuracy, precision, recall and F1-score equal to 92.20% due to its ability to capture more complex spatial features. This comparative analysis may guide the researchers in selecting appropriate models for real-world applications for action recognition. In addition of this, a new dataset is developed called Anomaly-5 that can helps as a valuable resource for training and evaluating action recognition algorithms.
|
| format | Article |
| id | doaj-art-ebab393eb94440ccb6b8de624e715a9e |
| institution | Kabale University |
| issn | 2520-0755 2522-3003 |
| language | English |
| publishDate | 2024-10-01 |
| publisher | Sukkur IBA University |
| record_format | Article |
| series | Sukkur IBA Journal of Computing and Mathematical Sciences |
| spelling | doaj-art-ebab393eb94440ccb6b8de624e715a9e2025-02-11T19:23:46ZengSukkur IBA UniversitySukkur IBA Journal of Computing and Mathematical Sciences2520-07552522-30032024-10-018110.30537/sjcms.v8i1.1444Comparative Analysis of Pre-trained based CNN-RNN Deep Learning Models on Anomaly-5 Dataset for Action Recognition Fayaz Ahmed Memon0Umair Ali Khan1Pardeep Kumar2Imtiaz Ali Halepoto3Farida Memon4Software Engineering Department, Quaid-e-Awam University of Engineering, Science & Technology, NawabshahArtificial Intelligence Department, QUEST NawabshahSoftware Engineering Department, Quaid-e-Awam University of Engineering, Science & Technology, Nawabshah.Department of Software Engineering, QUEST NawabshahDepartment of Electronic Engineering, Mehran University of Engineering & Technology, Jamshoro Action recognition in videos is one of the essential, challenging and active area of research in the field of computer vision that adopted in various applications including automated surveillance systems, security systems and human computer interaction. In this paper, we present an in-depth comparative analysis of five CNN-RNN models based on pre-trained networks such as InceptionV3, VGG16, MobileNetV2, ResNet152V2 and InceptionResNetV2 with recurrent LSTM units for action recognition on Anomaly-5 dataset. The performance of these models is analyzed and compared in terms of accuracy, precision, recall & F1-scores and computational efficiency. The CNN-RNN architectures we considered for analysis in this paper, the ResNet152V2 based CNN-RNN model exhibits better performance and achieved highest accuracy, precision, recall and F1-score equal to 92.20% due to its ability to capture more complex spatial features. This comparative analysis may guide the researchers in selecting appropriate models for real-world applications for action recognition. In addition of this, a new dataset is developed called Anomaly-5 that can helps as a valuable resource for training and evaluating action recognition algorithms. https://journal.iba-suk.edu.pk:8089/sibajournals/index.php/sjcms/article/view/1444 |
| spellingShingle | Fayaz Ahmed Memon Umair Ali Khan Pardeep Kumar Imtiaz Ali Halepoto Farida Memon Comparative Analysis of Pre-trained based CNN-RNN Deep Learning Models on Anomaly-5 Dataset for Action Recognition Sukkur IBA Journal of Computing and Mathematical Sciences |
| title | Comparative Analysis of Pre-trained based CNN-RNN Deep Learning Models on Anomaly-5 Dataset for Action Recognition |
| title_full | Comparative Analysis of Pre-trained based CNN-RNN Deep Learning Models on Anomaly-5 Dataset for Action Recognition |
| title_fullStr | Comparative Analysis of Pre-trained based CNN-RNN Deep Learning Models on Anomaly-5 Dataset for Action Recognition |
| title_full_unstemmed | Comparative Analysis of Pre-trained based CNN-RNN Deep Learning Models on Anomaly-5 Dataset for Action Recognition |
| title_short | Comparative Analysis of Pre-trained based CNN-RNN Deep Learning Models on Anomaly-5 Dataset for Action Recognition |
| title_sort | comparative analysis of pre trained based cnn rnn deep learning models on anomaly 5 dataset for action recognition |
| url | https://journal.iba-suk.edu.pk:8089/sibajournals/index.php/sjcms/article/view/1444 |
| work_keys_str_mv | AT fayazahmedmemon comparativeanalysisofpretrainedbasedcnnrnndeeplearningmodelsonanomaly5datasetforactionrecognition AT umairalikhan comparativeanalysisofpretrainedbasedcnnrnndeeplearningmodelsonanomaly5datasetforactionrecognition AT pardeepkumar comparativeanalysisofpretrainedbasedcnnrnndeeplearningmodelsonanomaly5datasetforactionrecognition AT imtiazalihalepoto comparativeanalysisofpretrainedbasedcnnrnndeeplearningmodelsonanomaly5datasetforactionrecognition AT faridamemon comparativeanalysisofpretrainedbasedcnnrnndeeplearningmodelsonanomaly5datasetforactionrecognition |