The goal of this paper is to find the query word in the dataset of images and to recognize the content of word image which is aided by the dictionary or lexicon. But in this paper there is the problem of word spotting and word recognition on images. In the proposed system, This is achieved by a combination of label embedding and attributes learning, and a common subspace regression. Then the images and strings represent the same word which are close to each other allowing one to cast recognition and retrieval tasks. Compared with the existing method, The advantage of our method has a fixed length , low dimensional and very fast to compute. In the preprocessing the given dataset is filtered by using median filter. After that , in the segmentation process every image is cropped identically. Then in the feature extraction is done by Gabor wavelet. For classifying the image we use KNN classifier. Matlab software and its image processing toolbox have been used in images processing and analysis. We test our approach for the given dataset of both handwritten documents and natural images showing results comparable or better than the state-of-the-art on spotting and recognition tasks.