Visual tracking with VG-RAM Weightless Neural Networks

Berger, Mariella; Souza, Alberto F. De; Neto, Jorcy de Oliveira; Aguiar, Edilson de; Oliveira-Santos, Thiago

doi:10.1016/j.neucom.2015.04.127

Cited by 8 publications

(4 citation statements)

References 30 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…However, latest works proved that adding small perturbations to the input are sufficient to mislead its conclusions [27]. In contrast, people are naturally capable to track variations, where AI still struggles in performing equivalent tasks despite the recent advances [28]. For example, some work exhibited that a one-pixel attack performed is enough to fool a network, using adversarial methods [29].…”

Section: Discussionmentioning

confidence: 99%

Novel Preprocessors for Convolution Neural Networks

et al. 2022

View full text Add to dashboard Cite

Fooling neural networks is a main concern in the process of Artificial Intelligence optimization. Character perturbation make part of a text unnoticeable for some systems, even for human observers. This research focuses the probe on a novel input preprocessing technique, which applies with the Convolutional Neural Networks for character recognition applications. Using Hand-written, Font, and Arabic Alphabet datasets, we show the CNN competitive results in classification. Then, we apply random deformation by mean of reshaping and relocation on the three testing datasets. As a reaction to change, the CNN demonstrates a confused behavior. Therefore, we design a Weightless Neural Network preprocessor that optimizes the shapes of deformed characters. Finally, we compare the impact of corrected input in improving the CNN performance. To measure the success of our preprocessor, we have applied it on well-known CNNs to compare the results between defected and rectified test datasets. Using Keras, we reconstructed LeNet-5, AlexNet and ZFNet for testing. Then, we conduct the test on the three databases to show how the recognition capability of CNNs would decline, then improve after preprocessing.INDEX TERMS Artificial neural networks, logic-based intelligence, pattern recognition, preprocessor, weightless neural networks.

show abstract

Section: Discussionmentioning

confidence: 99%

Novel Preprocessors for Convolution Neural Networks

et al. 2022

View full text Add to dashboard Cite

show abstract

“…The approach is taken on this face detection, such as lighting, position views, colours, shadows, noise, or image resolution [19]- [20]. Face tracking is used to follow faces from time to time in real-time by determining face changes (scale and position of faces) to allocate faces in a frame [15]- [16], [21]. This approach will easily be implemented on a computer system.…”

Section: Introductionmentioning

confidence: 99%

Robot movement controller based on dynamic facial pattern recognition

Nurmaini

Zarkasi

Stiawan

et al. 2021

IJEECS

View full text Add to dashboard Cite

In terms of movement, mobile robots are equipped with various navigation techniques. One of the navigation techniques used is facial pattern recognition. But Mobile robot hardware usually uses embedded platforms which have limited resources. In this study, a new navigation technique is proposed by combining a face detection system with a ram-based artificial neural network. This technique will divide the face detection area into five frame areas, namely top, bottom, right, left, and neutral. In this technique, the face detection area is divided into five frame areas, namely top, bottom, right, left, and neutral. The value of each detection area will be grouped into the ram discriminator. Then a training and testing process will be carried out to determine which detection value is closest to the true value, which value will be compared with the output value in the output pattern so that the winning discriminator is obtained which is used as the navigation value. In testing 63 face samples for the Upper and Lower frame areas, resulting in an accuracy rate of 95%, then for the Right and Left frame areas, the resulting accuracy rate is 93%. In the process of testing the ram-based neural network algorithm pattern, the efficiency of memory capacity in ram, the discriminator is 50%, assuming a 16-bit input pattern to 8 bits. While the execution time of the input vector until the winner of the class is under milliseconds (ms).

show abstract

“…To identify the most important sample for training an MIL tracker, the Pulse Coupled NN is adopted in [33] to identify the tracking target region. A long-term object tracking system is designed in [34] based on virtual generalizing random access memory weightless NNs. In this tracker, three constraints restrictions are imposed for updating the model.…”

Section: Introductionmentioning

confidence: 99%

Object Tracking Based on Compressive Features and Extreme Learning Machine

2019

View full text Add to dashboard Cite

Effective object tracking is a great challenging task in computer vision due to the appearance changes of an object. Extreme learning machine (ELM) can analytically determine the weights of feed-forward neural networks, while compressive features can be extracted via a random sparse matrix. Thus, their common advantages are running fast. Moreover, the strong learning ability of ELM is helpful to make a tracker robust target appearance change. In this paper, we propose an effective tracking method using ELM and compressive features. First, a bank of compressive Haar-like features is extracted with the sparse random matrix. To make the feature invariant to illumination, the mean illumination of each image patch is removed from each Haar-like feature. Second, a bank of ELM-based weak classifiers are combined together to form a strong classifier to complete tracking task. Considering the discrimination performance of features is uneven, the training accuracy of each ELM is employed as the confidence (weight) of the weak classifier. Thus, less discriminative features will be weakened. The proposed tracker has been tested on several representative sequences, and the results show that it achieves the highest tracking success rate and the tracking accuracy among the compared trackers including the FCT, CNT, MIL, KCF, and ESPT. INDEX TERMSExtreme learning machine, training accuracy, compressive features, object tracking.

show abstract

Visual tracking with VG-RAM Weightless Neural Networks

Cited by 8 publications

References 30 publications

Novel Preprocessors for Convolution Neural Networks

Novel Preprocessors for Convolution Neural Networks

Robot movement controller based on dynamic facial pattern recognition

Object Tracking Based on Compressive Features and Extreme Learning Machine

Contact Info

Product

Resources

About