Optimization of deep learning features for age-invariant face recognition

Abstract

This paper presents a methodology for Age-Invariant Face Recognition (AIFR), based on the optimization of deep learning features. The proposed method extracts deep learning features using transfer deep learning, extracted from the unprocessed face images. To optimize the extracted features, a Genetic Algorithm (GA) procedure is designed in order to select the most relevant features to the problem of identifying a person based on his/her facial images over different ages. For classification, K-Nearest Neighbor (KNN) classifiers with different distance metrics are investigated, i.e., Correlation, Euclidian, Cosine, and Manhattan distance metrics. Experimental results using a Manhattan distance KNN classifier achieves the best Rank-1 recognition rate of 86.2% and 96% on the standard FGNET and MORPH datasets, respectively. Compared to the state-of-the-art methods, our proposed method needs no preprocessing stages. In addition, the experiments show its privilege over other related methods.

Deep segmentation of the liver and the hepatic tumors from abdomen tomography images

“…We applied the two models (Densenet and FCN-Alexnet) in the proposed system, since their decoders produce outputs that are of the same dimensions as the input image, which suits the task of segmentation. In addition, they have shown outstanding performance for several related medical applications, such as lung segmentation [17], [18], pulmonary cancerous detection [19], face recognition [20], brain cancer [21] and diabetic retinopathy [22].…”

Pedestrian age estimation based on deep learning

“…Both ResNet-50 convolution network model suggested in [22], and VGG-Face convolution network model, suggested in [23], [24] have been used in this work to achieve some of the best performance in an age estimation task. ResNet-50 is a deep CNNs based on residual neuronal network architecture.…”

“…Although the black and white face images are robust in face recognition but they are excluded, The VGGFace2 dataset focused on facial and image variation due to color processing as shown in Figure 4. Five age classes have been included in this study {(00-10), (11)(12)(13)(14)(15)(16)(17)(18)(19)(20), (21)(22)(23)(24)(25)(26)(27)(28)(29)(30)(31)(32)(33)(34)(35), (36-54), (55-90)}.…”

“…Before loading input images, the working environment need some preparation such as installing the required libraries, then start to load the dataset. A preprocessing data, is required at this stage including cleansing the dataset from low quality images which will confuse the training model, A relabeling age group images to five categories {(00-10), (11)(12)(13)(14)(15)(16)(17)(18)(19)(20), (21)(22)(23)(24)(25)(26)(27)(28)(29)(30)(31)(32)(33)(34)(35), (36-54), (55-90)}, splitting dataset to 80% train, 10% test and 10% validation and finally, set the batch size. The age estimation model is trained with three pre-trained-weight, and two deep learning algorithms (VGG-Face and ResNet50) as it illustrated in the experimental work.…”

An image-based gangrene disease classification

“…The CNN is widely used in feature extraction from the images. One of the recent works [18] proposes the feature extraction of face images to do the age-invariant face recognition.…”