Sketch recognition

About: Sketch recognition is a research topic. Over the lifetime, 1611 publications have been published within this topic receiving 40284 citations.

Hierarchical primitive shape classification based on cascade feature point detection for early processing of on-line sketch recognition

Abstract: A new approach for on-line primitive shape classification for early processing of sketch recognition is proposed in this paper. Firstly, cascade feature point detection composed of three steps is implemented, which can remove noise effectively and retain the characteristic of the shape. Then a hierarchical primitive shape classification method is applied which has an active effect on discrimination. Sketch input is firstly determined as a close or non-close shape through the close/non-close shape judgement. Following that two SVM classifiers are utilized to determine the exact type with a combined feature based on velocity and turning angle being adopted to improve recognition accuracy. Finally, the experiments prove the effectiveness and efficiency of the proposed approach in robust early processing for freehand sketching.

Learning Through the Lens of Sketch

Abstract: Dr. Tracy Hammond gave a keynote on morning of the second day of the conference. She spoke about how both her research and the field of Sketch Recognition evolved over the last decade. One motivation of her career was to develop algorithms that provide insights into human brain activity and also develop applications that improve human-human communication. Her initial work focused on domain-independent recognition methods, while her current work focuses on developing systems to improve education. She also shows how sketch recognition methods can advance both sketch forensics and activity recognition, providing inspiration as per how this can allow for surprisingly intelligent personalized feedback. This chapter provides a lightly edited transcription of that keynote.

Method of determining training data for gesture recognition considering decay in gesture movements

Abstract: Mobile phones and video game controllers using gesture recognition technologies enable easy and intuitive operations, such as those in drawing objects. Gesture recognition systems generally require several samples of training data before recognition takes place. However, recognition accuracy deteriorates as time passes since the trajectory of the gestures changes due to fatigue or forgetfulness. We investigated the change in gestures and fast found that several samples of gestures were not suitable for training data. Therefore, we propose two methods of finding appropriate data for training. We confirmed that the proposed methods found better training data than the conventional method from the viewpoints of the number of data collected and the accuracy of recognition.

Utilizing temporal associations for view-based 3-D object recognition

Abstract: The authors propose an architecture for the recognition of three-dimensional objects on the basis of viewer-centered representations and temporal associations. Motivated by biological findings and by successful computational implementations they have chosen a viewer-centered representation scheme. In contrast to other implementations, special attention is paid to the temporal order of the views, which proves useful for learning and recognition purposes. Their recognition system combines different kinds of artificial neural networks into a four stage architecture: preprocessing by a Gaborjet transform is followed by an extended dynamic link matching algorithm which implements recognition and learning of the view classes. A STORE network records the temporal order of the views by transforming a sequence of view classes into an item-and-order coding. Subsequently, a Gaussian-ARTMAP architecture is used for the classification of the sequences and for their mapping onto object classes by means of supervised learning. The presented results demonstrate that the system is capable to autonomously learn and to discriminate similar objects. Additionally, the examples show how the utilization of the temporal context improves object recognition by making ambiguous views manageable and facilitating an increased insensitiveness against misclassifications.

Computer vision for face-to-face human-computer interaction

Abstract: Computer vision will play an important role in future humancomputer interactions. In order to achieve natural humancomputer interaction, there is a need for the computer to be able to interact with the user similar to the way human-human interaction takes place. In recent years we have been investigating various methods for different tasks of face-to-face interaction. Many of these are using image and video descriptors based on local binary patterns (LBP). This talk overviews our recent progress in LBP descriptors, face recognition and biometrics, recognition of facial expressions and emotions, visual speech recognition, synthesis of a talking face, and recognition of human actions and gestures. Results of our research on an experimental system for intelligent humanrobot interaction are also discussed.