A contribution to semantic description of images and videos: an application of soft biometrics

Humans have a high ability to extract visual data information acquired by sight. Trought a learning process, which starts at birth and continues throughout life, image interpretation becomes almost instinctively. At a glance, one can easily describe a scene with reasonable precision, naming its main components. Usually, this is done by extracting low-level features such as edges, shapes and textures, and associanting them to high level meanings. In this way, a semantic description of the scene is done. An example of this, is the human capacity to recognize and describe other people physical and behavioral characteristics, or biometrics. Soft-biometrics also represents inherent characteristics of human body and behaviour, but do not allow unique person identification. Computer vision area aims to develop methods capable of performing visual interpretation with performance similar to humans. This thesis aims to propose computer vison methods which allows high level information extraction from images in the form of soft biometrics. This problem is approached in two ways, unsupervised and supervised learning methods. The first seeks to group images via an automatic feature extraction learning , using both convolution techniques, evolutionary computing and clustering. In this approach employed images contains faces and people. Second approach employs convolutional neural networks, which have the ability to operate on raw images, learning both feature extraction and classification processes. Here, images are classified according to gender and clothes, divided into upper and lower parts of human body. First approach, when tested with different image datasets obtained an accuracy of approximately 80% for faces and non-faces and 70% for people and non-person. The second tested using images and videos, obtained an accuracy of about 70% for gender, 80% to the upper clothes and 90% to lower clothes. The results of these case studies, show that proposed methods are promising, allowing the realization of automatic high level information image annotation. This opens possibilities for development of applications in diverse areas such as content-based image and video search and automatica video survaillance, reducing human effort in the task of manual annotation and monitoring.

Proposta de um processo de transferência de tecnologia para as instituições públicas de pesquisa: o caso da Embrapa

The Brazilian agricultural research agency has, over the years, contributed to solve social problems and to promote new knowledge, incorporating new advances and seeking technological independence of the country, through the transfer of knowledge and technology generated. However, the process of transfering of knowledge and technology has represented a big challenge for public institutions. The Embrapa is the largest and main brazilian agricultural research company, with a staff of 9.790 employees, being 2.440 researchers and an annual budget of R$ 2.52 billion. Operates through 46 decentralized research units, and coordinate of the National Agricultural Research System - SNPA. Considering that technology transfer is the consecration of effort and resources spent for the generation of knowledge and the validity of the research, this work aims to conduct an assessment of the performance of Embrapa Swine and Poultry along the production chain of broilers and propose a technology transfer model for this chain, which can be used by the Public Institutions Research – IPPs. This study is justified by the importance of agricultural research for the country, and the importance of the institution addressed. The methodology used was the case study with a qualitative approach, documentary and bibliographic research and interviews with use of semi-structured questionnaires. The survey was conducted in three stages. In the first stage, there was a diagnosis of the Technology Transfer Process (TT), the contribution of the Embrapa Swine and poultry for the supply chain for broiler. At this stage it was used bibliographical and documentary research and semi- structured interviews with agroindustrial broiler agents, researchers at Embrapa Swine and Poultry, professionals of technology transfer, from the Embrapa and Embrapa Swine and Poultry, managers of technology transfer and researchers from the Agricultural Research Service - ARS. In the second step, a model was developed for the technology transferring poultry process of Embrapa. In this phase, there were made documentary and bibliographic research and analysis of information obtained in the interviews. The third phase was to validate the proposed model in the various sectors of the broilers productive chain. The data show that, although the Embrapa Swine and Poultry develops technologies for broiler production chain, the rate of adoption of these technologies by the chain is very low. It was also diagnosed that there is a gap between the institution and the various links of the chain. It was proposed an observatory mechanism to approximate Embrapa Swine and Poultry and the agents of the broiler chain for identifying and discussing research priorities. The proposed model seeks to improve the interaction between the institution and the chain, in order to identify the chain real research demands and the search and the joint development of solutions for these demands. The proposed TT model was approved by a large majority (96.77%) of the interviewed agents who work in the various links in the chain, as well as by representatives (92%) of the entities linked to this chain. The acceptance of the proposed model demonstrates the willingness of the chain to approach Embrapa Swine and Poultry, and to seek joint solutions to existing problems.