Generalized zero-shot action recognition is a challenging problem, where the task is to recognize new action categories that are unavailable during the training stage, in addition to the seen action categories. Existing approaches suffer from the inherent bias of the learned classifier towards the seen action categories. As a consequence, unseen category samples are incorrectly classified as belonging to one of the seen action categories. In this paper, we set out to tackle this issue by arguing for a separate treatment of seen and unseen action categories in generalized zeroshot action recognition. We introduce an out-of-distribution detector that determines whether the video features belong to a seen or unseen action category. To train our out-ofdistribution detector, video features for unseen action categories are synthesized using generative adversarial networks trained on seen action category features. To the best of our knowledge, we are the first to propose an outof-distribution detector based GZSL framework for action recognition in videos. Experiments are performed on three action recognition datasets: Olympic Sports, HMDB51 and UCF101. For generalized zero-shot action recognition, our proposed approach outperforms the baseline [33] with absolute gains (in classification accuracy) of 7.0%, 3.4%, and 4.9%, respectively, on these datasets. * Authors contributed equally.

No abstract

No abstract

Cross-modal retrieval is gaining importance due to the availability of large amounts of multimedia data. Hashing-based techniques provide an attractive solution to this problem when the data size is large. For cross-modal retrieval, data from the two modalities may be associated with a single label or multiple labels, and in addition, may or may not have a one-to-one correspondence. This work proposes a simple hashing framework which has the capability to work with different scenarios while effectively capturing the semantic relationship between the data items. The work proceeds in two stages in which the first stage learns the optimum hash codes by factorizing an affinity matrix, constructed using the label information. In the second stage, ridge regression and kernel logistic regression is used to learn the hash functions for mapping the input data to the bit domain. We also propose a novel iterative solution for cases where the training data is very large, or when the whole training data is not available at once. Extensive experiments on single label data set like Wiki and multi-label datasets like MirFlickr, NUS-WIDE, Pascal, and LabelMe, and comparisons with the state-of-the-art, shows the usefulness of the proposed approach.

This study was conducted on 94 Frieswal (5/8 Holstein Friesian 3/8 Sahiwal) crossbred bulls of three different grades, categorized based on their semen freezability visualising Group 1 (consistently freezable semen producer bulls, N 5 11), Group 2 (inconsistent freezable, N 5 16) and Group 3 (Non freezable, N 5 67). Each group was further divided into two classes that is young (up to 30 months) and adult (31 to 70 months) bulls depending upon their age. Sperm morphology was studied by using the eosin-nigrosin staining technique. Bulls age significantly (P , 0.01) affected semen quality and sperm morphology. In adult bulls, semen volume, mass activity and sperm concentration were 36%, 17.56% and 19.6%, respectively, higher than young. Initial progressive motility (%) and livability showed significant (P , 0.01) improvement with the advancement of age (43.37 6 1.21 and 67.71 6 1.11, respectively, in young; 53.02 6 1.11 and 74.17 6 1.03, respectively, in adult). In young bulls, sperm head, mid piece, tail abnormality and total abnormal sperm percent (12.38 6 0.92, 4.87 6 0.24, 11.01 6 0.60 and 28.26 6 1.34, respectively) were 1.85, 1.27, 1.20 and 1.44 folds higher than that of their mature stage (6.69 6 0.64, 3.82 6 0.32, 9.14 6 0.64 and 19.66 6 1.31, respectively). Significant reduction (P , 0.01) in micro cephalic sperm, free heads, bent mid piece, looped mid piece and proximal protoplasmic droplets were observed at mature age as compared with their younger stage. In bulls of consistent freezing category, abnormal sperm heads significantly decreased from 4.40 6 0.31% to 3.28 6 0.02% on maturity. Similarly, in inconsistent freezing grade bulls sperm head abnormality (9.28 6 0.75% to 5.13 6 1.20%) and total abnormal sperm percent (24.89 6 1.43 to 18.73 6 3.40) was decreased over the age. On the contrary, in non-freezing category bulls' sperm morphology did not show significant (P . 0.05) improvement with age advancement, rather some abnormalities like long slender head, under developed/deformed head, abaxial implantation of mid piece, double mid piece, stump tail and distal protoplasmic droplets tend to increased significantly (P , 0.05) with age of bulls. Results indicated that in potential Frieswal bulls semen quality and sperm morphology were improved from young to mature stage, where as, in poor quality (non-freezing) semen producer bulls neither the morphology nor the semen quality showed any improvement with maturity. It was recommended that crossbred bulls producing more than 25% morphologically abnormal sperms in young age (below 30 months) along with poor progressive motility (,50%) and low sperm concentration (,1000 million/ml) need immediate culling with out any expectation of further improvement in semen quality with age advancement.Keywords: Sperm abnormality, Frieswal crossbred bull, Sperm morphology, Bos indicus 3 Bos taurus Implication Assessment of sperm morphology is one of the most important aspects for semen evaluation and successful fertilisation. High percentage of morphologically abnormal spe...

No abstract

No abstract

No abstract