Deep learning for face detection using matlab
This project report presents face detection using Convolutional Neural Network algorithm and Deep Learning combination (DCT / DL) throughout MATLAB simulation and modeling. It reveals that the research project has successfully managed to establish an accurate accurate human face detection and crysta...
| Main Author: | |
|---|---|
| Format: | Thesis |
| Language: | English English English |
| Published: |
2020
|
| Subjects: | |
| Online Access: | http://eprints.uthm.edu.my/398/1/24p%20SALIM%20ADNAN%20SALIM.pdf http://eprints.uthm.edu.my/398/2/SALIM%20ADNAN%20SALIM%20COPYRIGHT%20DECLARATION.pdf http://eprints.uthm.edu.my/398/3/SALIM%20ADNAN%20SALIM%20WATERMARK.pdf |
| _version_ | 1825636462993866752 |
|---|---|
| author | Slim, Salim Adnan |
| author_facet | Slim, Salim Adnan |
| author_sort | Slim, Salim Adnan |
| collection | UTHM |
| description | This project report presents face detection using Convolutional Neural Network algorithm and Deep Learning combination (DCT / DL) throughout MATLAB simulation and modeling. It reveals that the research project has successfully managed to establish an accurate accurate human face detection and crystal-clear human face recognition systems. The system will annul the face image that are tilted, the images on non-human faces as well as the images of human faces that have watermarks. The test results on face tracking when the image has watermarks. Under this condition, it looks like the CNN and deep learning could not identify the image correctly and wrong result is showing for the second image. This indicates that there is a limitation for the CNN and deep learning algorithm. The disadvantage is, it cannot detects the watermark image, as this image is protected. This process will proceed to Convolutional Neural Network algorithm to identify the human face from a given image. If the image belongs to human features then there will be a tracking box marking clearly the appointed face with a yellow square. This marker will be clearly pointing and shaping a yellowish box of the appointed and selected face or image.The novelty of this research project is that the CNN and deep learning (CNN / DL) methods to trace, scan and detect the human face in a very successful and effective manners taking into account the following distinguished features: the face of the human facing to the front view and not tilted or the face does not make any angles unless angels within 5 to 10 degrees only. The face must not be hiding or nor recognizable or positioned on another object. The face must be real and is not printed on any object like wood or plastic. The water mark must not be printed on the face image picture, otherwise the CNN / DL will not recognize it as human face. The working of the algorithm depends on the deep learning where the system needs to learn the image, identify the faces and store the images into database. By creating a folder called image folder, it will be easy for the MATLAB access into the folder to find the images that content human face and none human face. High resolution for face detection was approximately 85%. The algorithm was able to distinguish between human and non-human faces. By doing this we saved a lot of time in almost half. |
| first_indexed | 2024-03-05T21:37:13Z |
| format | Thesis |
| id | uthm.eprints-398 |
| institution | Universiti Tun Hussein Onn Malaysia |
| language | English English English |
| last_indexed | 2024-03-05T21:37:13Z |
| publishDate | 2020 |
| record_format | dspace |
| spelling | uthm.eprints-3982021-07-25T02:20:22Z http://eprints.uthm.edu.my/398/ Deep learning for face detection using matlab Slim, Salim Adnan Q300-390 Cybernetics This project report presents face detection using Convolutional Neural Network algorithm and Deep Learning combination (DCT / DL) throughout MATLAB simulation and modeling. It reveals that the research project has successfully managed to establish an accurate accurate human face detection and crystal-clear human face recognition systems. The system will annul the face image that are tilted, the images on non-human faces as well as the images of human faces that have watermarks. The test results on face tracking when the image has watermarks. Under this condition, it looks like the CNN and deep learning could not identify the image correctly and wrong result is showing for the second image. This indicates that there is a limitation for the CNN and deep learning algorithm. The disadvantage is, it cannot detects the watermark image, as this image is protected. This process will proceed to Convolutional Neural Network algorithm to identify the human face from a given image. If the image belongs to human features then there will be a tracking box marking clearly the appointed face with a yellow square. This marker will be clearly pointing and shaping a yellowish box of the appointed and selected face or image.The novelty of this research project is that the CNN and deep learning (CNN / DL) methods to trace, scan and detect the human face in a very successful and effective manners taking into account the following distinguished features: the face of the human facing to the front view and not tilted or the face does not make any angles unless angels within 5 to 10 degrees only. The face must not be hiding or nor recognizable or positioned on another object. The face must be real and is not printed on any object like wood or plastic. The water mark must not be printed on the face image picture, otherwise the CNN / DL will not recognize it as human face. The working of the algorithm depends on the deep learning where the system needs to learn the image, identify the faces and store the images into database. By creating a folder called image folder, it will be easy for the MATLAB access into the folder to find the images that content human face and none human face. High resolution for face detection was approximately 85%. The algorithm was able to distinguish between human and non-human faces. By doing this we saved a lot of time in almost half. 2020-01 Thesis NonPeerReviewed text en http://eprints.uthm.edu.my/398/1/24p%20SALIM%20ADNAN%20SALIM.pdf text en http://eprints.uthm.edu.my/398/2/SALIM%20ADNAN%20SALIM%20COPYRIGHT%20DECLARATION.pdf text en http://eprints.uthm.edu.my/398/3/SALIM%20ADNAN%20SALIM%20WATERMARK.pdf Slim, Salim Adnan (2020) Deep learning for face detection using matlab. Masters thesis, Universiti Tun Hussein Onn Malaysia. |
| spellingShingle | Q300-390 Cybernetics Slim, Salim Adnan Deep learning for face detection using matlab |
| title | Deep learning for face detection using matlab |
| title_full | Deep learning for face detection using matlab |
| title_fullStr | Deep learning for face detection using matlab |
| title_full_unstemmed | Deep learning for face detection using matlab |
| title_short | Deep learning for face detection using matlab |
| title_sort | deep learning for face detection using matlab |
| topic | Q300-390 Cybernetics |
| url | http://eprints.uthm.edu.my/398/1/24p%20SALIM%20ADNAN%20SALIM.pdf http://eprints.uthm.edu.my/398/2/SALIM%20ADNAN%20SALIM%20COPYRIGHT%20DECLARATION.pdf http://eprints.uthm.edu.my/398/3/SALIM%20ADNAN%20SALIM%20WATERMARK.pdf |
| work_keys_str_mv | AT slimsalimadnan deeplearningforfacedetectionusingmatlab |