Modified Convolutional Neural Network Architecture for Batik Motif Image Classification

In this paper, we propose a learning-based supervised discrete hashing (SDH) method. Binary hashing is widely used for large-scale image retrieval as well as video and document searches, because the compact binary code representation is essential for data storage and reasonable for query searches using bit operations. The recently proposed SDH method efficiently solves mixed-integer programming problems by alternating optimization and the discrete cyclic coordinate descent (DCC) method. Based on some preliminary experiments, we show that the SDH method can be simplified without performance degradation. We analyze the simplified model and provide a mathematically exact solution thereof; we reveal that the exact binary code is provided by a “Hadamard matrix.” Therefore, we named our method Hadamard coded-SDH (HC-SDH). In contrast to the SDH, our model does not require an alternating optimization algorithm and does not depend on initial values. The HC-SDH is also easier to implement than the iterative quantization. Experimental results involving a large-scale database show that the Hadamard coding outperforms the conventional SDH in terms of precision, recall, and computational time. On the large data sets SUN-397 and ImageNet, the HC-SDH provides a superior mean average of precision (mAP) and top-accuracy compared with the conventional SDH methods with the same code length and FastHash. The training time of the HC-SDH is 170 times faster than the conventional SDH and the testing time including the encoding time is seven times faster than the FastHash which encodes using a binary-tree.

Grasping and manipulating transparent objects with a robot is a challenge in robot vision. To successfully perform robotic grasping, 6D object pose estimation is needed. However, transparent objects are difficult to recognize because their appearance varies depending on the background, and modern 3D sensors cannot collect reliable depth data on transparent object surfaces due to the translucent, refractive, and specular surfaces. To address these challenges, we proposed a 6D pose estimation of transparent objects for manipulation. Given a single RGB image of transparent objects, the 2D keypoints are estimated using a deep neural network. Then, the PnP algorithm takes camera intrinsics, object model size, and keypoints as inputs to estimate the 6D pose of the object. Finally, the predicted poses of the transparent object were used for grasp planning. Our experiments demonstrated that our picking system is capable of grasping transparent objects from different backgrounds. To the best of our knowledge, this is the first time a robot has grasped transparent objects from a single RGB image. Furthermore, the experiments show that our method is better than the 6D pose estimation baselines and can be generalized to real-world images.

In recent years, many researchers have shown increasing interest in music information retrieval (MIR) applications, with automatic chord recognition being one of the popular tasks. Many studies have achieved/demonstrated considerable improvement using deep learning based models in automatic chord recognition problems. However, most of the existing models have focused on simple chord recognition, which classifies the root note with the major, minor, and seventh chords. Furthermore, in learning-based recognition, it is critical to collect high-quality and large amounts of training data to achieve the desired performance. In this paper, we present a multi-task learning (MTL) model for a guitar chord recognition task, where the model is trained using a relatively large-vocabulary guitar chord dataset. To solve data scarcity issues, a physical data augmentation method that directly records the chord dataset from a robotic performer is employed. Deep learning based MTL is proposed to improve the performance of automatic chord recognition with the proposed physical data augmentation dataset. The proposed MTL model is compared with four baseline models and its corresponding single-task learning model using two types of datasets, including a human dataset and a human combined with the augmented dataset. The proposed methods outperform the baseline models, and the results show that most scores of the proposed multi-task learning model are better than those of the corresponding single-task learning model. The experimental results demonstrate that physical data augmentation is an effective method for increasing the dataset size for guitar chord recognition tasks.

When learning to play a musical instrument, it is important to improve the quality of self-practice. Many systems have been developed to assist practice. Some practice assistance systems use special sensors (pressure, flow, and motion sensors) to acquire the control parameters of the musical instrument, and provide specific guidance. However, it is difficult to acquire the control parameters of wind instruments (e.g., saxophone or flute) such as flow and angle between the player and the musical instrument, since it is not possible to place sensors into the mouth. In this paper, we propose a sensorless control parameter estimation system based on the recorded sound of a wind instrument using only machine learning. In the machine learning framework, many training samples that have both sound and correct labels are required. Therefore, we generated training samples using a robotic performer. This has two advantages: (1) it is easy to obtain many training samples with exhaustive control parameters, and (2) we can use the correct labels as the given control parameters of the robot. In addition to the samples generated by the robot, some human performance data were also used for training to construct an estimation model that enhanced the feature differences between robot and human performance. Finally, a flute control parameter estimation system was developed, and its estimation accuracy for eight novice flute players was evaluated using the Spearman’s rank correlation coefficient. The experimental results showed that the proposed system was able to estimate human control parameters with high accuracy.

This paper presents Linear Mean-Median (LMM) filter that used to reduce impulse noise. LMM filter is a combination between Mean and Median filter. Wherein, linear value is acquired from the linearity between mean and median value. Mean and Median filter are only applied for free-noise pixel on the 3×3 windows that has been sorted from the smallest to the largest value. The mean value is obtained from the average value of all free-noise pixels without including the median pixel position. Meanwhile, median pixel is the middle position of the pixel that has been sorted. LMM uses nine sample pixels to determine a pixel for replacement a corrupted pixel. Our filter also provides the impulse noise prediction systems that serve as a facilitator to give information about noise content. If the noise is greater than 30%, the performance of LMM filter needs to be improved by an adaptive rank order mean filters. The filtering results have shown satisfactory results in terms of the quality result and the computation time process. A good image quality can be evidenced by PSNR (Peak Signal to Noise Ratio). Our methods always have higher PSNR value than the comparison methods. In addition, the speed computation time of our method is faster than the comparison method.