Deep Convolutional Neural Networks (DCNN) have established a remarkable performance benchmark in the field of image classification, displacing classical approaches based on hand-tailored aggregations of local descriptors. Yet DCNNs impose high computational burdens both at training and at testing time, and training them requires collecting and annotating large amounts of training data. Supervised adaptation methods have been proposed in the literature that partially re-learn a transferred DCNN structure from a new target dataset. Yet these require expensive bounding-box annotations and are still computationally expensive to learn. In this paper, we address these shortcomings of DCNN adaptation schemes by proposing a hybrid approach that combines conventional, unsupervised aggregators such as Bag-of-Words (BoW), with the DCNN pipeline by treating the output of intermediate layers as densely extracted local descriptors.We test a variant of our approach that uses only intermediate DCNN layers on the standard PASCAL VOC 2007 dataset and show performance significantly higher than the standard BoW model and comparable to Fisher vector aggregation but with a feature that is 150 times smaller. A second variant of our approach that includes the fully connected DCNN layers significantly outperforms Fisher vector schemes and performs comparably to DCNN approaches adapted to Pascal VOC 2007, yet at only a small fraction of the training and testing cost.

No abstract

Facial expression is a major area for non-verbal language in day to day life communication. As the statistical analysis shows only 7 percent of the message in communication was covered in verbal communication while 55 percent transmitted by facial expression. Emotional expression has been a research subject of physiology since Darwin’s work on emotional expression in the 19th century. According to Psychological theory the classification of human emotion is classified majorly into six emotions: happiness, fear, anger, surprise, disgust, and sadness. Facial expressions which involve the emotions and the nature of speech play a foremost role in expressing these emotions. Thereafter, researchers developed a system based on Anatomic of face named Facial Action Coding System (FACS) in 1970. Ever since the development of FACS there is a rapid progress of research in the domain of emotion recognition. This work is intended to give a thorough comparative analysis of the various techniques and methods that were applied to recognize and identify human emotions. This analysis results will help to identify the proper and suitable techniques, algorithms and the methodologies for future research directions. In this paper extensive analysis on the various recognition techniques used to identify the complexity in recognizing the facial expression is presented. This work will also help researchers and scholars to ease out the problem in choosing the techniques used in the identification of the facial expression domain.