Detection of a human face and face identification is one of the most intensive areas under computer vision. It has unique significance in criminal detection, video surveillance and person re-identification. The application areas of face retrieval include security and law enforcement, commercial applications, and public information gathering. The recognition of a face image in real-time is still challenging due to various constraints such as under or over-exposed illumination, the pose and angle of the face, and the quality of the images employed for the training. Image indexing and the retrieval of most similar images for a given input image required more robust image features. In this work we tried to derive the prominent texture features from the face images using Gray Level Co-occurrence Matrix (GLCM) with Local Ternary Pattern (LTP) and Principal Component Analysis (PCA) applied to reduce size of the feature vector. The GLCM is used to find the co-occurrences of the pixel values in four directions i.e. 0, 45, 90, and 135 degrees angles in the upper and lower LTP images. This work proposes a novel feature extraction technique referred to as CoALTP and PCA feature. Proposed model is tested using ORL face database for face image retrieval. We found a substantial improvement in the retrieval rate in comparison to other methods.