Besides working with speech recognition and assistive technology, lately I have been working with some other research and applications of Digital Signal Processing theory:
On Informed Coding and Host Rejection for Communication over Inkjet Print-and-Scan Channels
Joceli Mayer and Steven J. Simske
This paper describes novel approaches to achieve robust communication over inkjet print-and-scan (IPS) color channels. The IPS color channel poses even greater challenges than the laser printer-and-scan channel due to the resulting mixing and spreading of the ink dots. We propose a novel informed coding and two host color rejection approaches, one based on a novel color rejection and another on a whitening filter, to deal with the aforementioned inkjet printer distortions. A substitutive spatial domain embedding is proposed to enable robustness optimization using the proposed informed coding. Analyses and examples are provided to evaluate the performance enhancement on robustness and transparency achievable by the proposed approaches.
Increasing the capacity of color print codes for robust communication over inkjet print-scan channels
Joceli Mayer
This investigation on robust and high capacity print codes aims to increase information payload in a given printed page area while providing robustness to channel errors including distortions originated by the inkjet printing and scanning processes. The approach includes statistical print-and-scan channel characterization, designing of robust segmentation using visual cues, unsupervised Bayesian color classification with expectation-maximization algorithm for parameters estimation of a mixture of Gaussians model and design of error correction codes. Results illustrate the performance evaluated under real channel and distortions conditions. High payload of 4592 bytes per squared inch is achieved with a robustness of 92\% to distortions due to the print-and-scan channel. Adding high-density information to printed materials enables interesting hardcopy document applications involving security, authentication, physical-electronic round tripping, item-level tagging, and consumer/product interaction.
Method for Correction of Lenses Distortion in Stereo Vision
Joceli Mayer and Osmando Pereira Junior
Lenses distortion is one of the main factors that limits the accuracy of stereo vision system reconstruction. We propose a new method for correction of the lenses distortion by applying compensation to each region of an image. Our method splits the image into smaller regions and compensates for each region for a fixed lenses model order. When compared to the conventional method, which models the entire image with only one model, our approach provides considerably better compensation and reduce the depth error as shown in the experiments with synthetic data.
Continuous Gesture Recognition using Hidden Markov Models
Joceli Mayer and Vinicius Breda
This work presents an algorithm for recognizing gestures in videos where the actions are executed continuously without pause between them and can be performed with one or both hands. We employ Hidden Markov Models (HMM) for modeling gestures as this technique has been applied successfully in speech and character recognition, We investigate the performance for a set of 26 visual descriptors extracted from the hands after a region segmentation based on normalized quadrants. Recognition are performed by adapting the Hidden Markov Models Toolkit (HTK) and achieve a recognition rate of 91.28% for a set of 21 phrases each composed of 4 gestures from a dictionary of 15 gestures from the Brazilian sign language (LIBRAS).
Informed communication system designed to embed hidden information into audio advertisements
Joceli Mayer and Luiz F.L.O. Silva
Investigation on informed communication system designed to embed hidden information into audio advertisements. The information is modulated and transmitted through a audio digital watermark embedded into the audio signal host. The watermark embedding is designed to achieve minimal perceptual impact and high robustness to environment distortions in order to comply with the proposed application in marketing. The embedding energy is properly scaled in order to achieve high transparency. Novel synchronization and adaptive equalization techniques are investigated and proposed to reliably extract the information after the channel interferences. Experimental results using simulated and real environment channels illustrate the performance of the proposed system and results are contrasted to other existing techniques.
Joceli Mayer 