|
|
Deep learning based face recognition in real conditions
Horňáková, Veronika ; Kříž, Petr (referee) ; Přinosil, Jiří (advisor)
This bachelor thesis explores the area of face recognition using deep learning technique. Face recognition is used for two main reasons: verification and identification. In this thesis we describe the techniques of deep learning, mostly the convolutional neural networks, which are the most significant method for processing images - detection, classification and segmentation of the image. The process of face recognition is divided into four main steps: face detection, face selection, face extraction and face classification. We chosen three of the existing programs for face recognition (OpenFace, FaceNet and Face_Recognition), which are described in this thesis, in particular the principle of the human face recognition. Thanks to the tests with the data set of Labeled Faces in the Wild (LFW) we could specify the accuracy and the time requirement of each application. Testing of FaceNet and Face_Recognition ran on real data with face detection in video with complicated conditions. The test compares two images and tries to determine if is the same person. The test results are show in graph and table.
|
|
|
The effect of the background and dataset size on training of neural networks for image classification
Mikulec, Vojtěch ; Kolařík, Martin (referee) ; Rajnoha, Martin (advisor)
This bachelor thesis deals with the impact of background and database size on training of neural networks for image classification. The work describes techniques of image processing using convolutional neural networks and the influence of background (noise) and database size on training. The work proposes methods which can be used to achieve faster and more accurate training process of convolutional neural networks. A binary classification of Labeled Faces in the Wild dataset is selected where the background is modified with color change or cropping for each experiment. The size of dataset is crucial for training convolutional neural networks, there are experiments with the size of training set in this work, which simulate a real problem with the lack of data when training convolutional neural networks for image classification.
|
|
|
The effect of the background and dataset size on training of neural networks for image classification
Mikulec, Vojtěch ; Kolařík, Martin (referee) ; Rajnoha, Martin (advisor)
This bachelor thesis deals with the impact of background and database size on training of neural networks for image classification. The work describes techniques of image processing using convolutional neural networks and the influence of background (noise) and database size on training. The work proposes methods which can be used to achieve faster and more accurate training process of convolutional neural networks. A binary classification of Labeled Faces in the Wild dataset is selected where the background is modified with color change or cropping for each experiment. The size of dataset is crucial for training convolutional neural networks, there are experiments with the size of training set in this work, which simulate a real problem with the lack of data when training convolutional neural networks for image classification.
|
|
|
Deep learning based face recognition in real conditions
Horňáková, Veronika ; Kříž, Petr (referee) ; Přinosil, Jiří (advisor)
This bachelor thesis explores the area of face recognition using deep learning technique. Face recognition is used for two main reasons: verification and identification. In this thesis we describe the techniques of deep learning, mostly the convolutional neural networks, which are the most significant method for processing images - detection, classification and segmentation of the image. The process of face recognition is divided into four main steps: face detection, face selection, face extraction and face classification. We chosen three of the existing programs for face recognition (OpenFace, FaceNet and Face_Recognition), which are described in this thesis, in particular the principle of the human face recognition. Thanks to the tests with the data set of Labeled Faces in the Wild (LFW) we could specify the accuracy and the time requirement of each application. Testing of FaceNet and Face_Recognition ran on real data with face detection in video with complicated conditions. The test compares two images and tries to determine if is the same person. The test results are show in graph and table.
|
guest ::
