Improving the lenet with batch normalization and online hard example mining for digits recognition

Xie, Yiliang; Huang, Jin; Tsang, Eric W. K.

doi:10.1109/icwapr.2017.8076680

Cited by 5 publications

(3 citation statements)

References 4 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Some authors presented a modified version of the LeNet network, for example, Lin et al [53] used smaller convolutional kernels (3 × 3) to increase the number of extracted features and reduced the fully connected layer from 10 units to 2 units. Xie et al [54] further improved LeNet by adding activation layers, batch normalisation layers, and online hard example mining. Li et al [55] increased the number of convolutions kernels at some of the layers, adopted the ReLu activation function, used max-pooling instead of mean-pooling layers, and used SVM at the output layer.…”

Section: Techniquesmentioning

confidence: 99%

Pedestrian and Vehicle Detection in Autonomous Vehicle Perception Systems—A Review

Galvao

Abbod

Kalganova

et al. 2021

Sensors

View full text Add to dashboard Cite

Autonomous Vehicles (AVs) have the potential to solve many traffic problems, such as accidents, congestion and pollution. However, there are still challenges to overcome, for instance, AVs need to accurately perceive their environment to safely navigate in busy urban scenarios. The aim of this paper is to review recent articles on computer vision techniques that can be used to build an AV perception system. AV perception systems need to accurately detect non-static objects and predict their behaviour, as well as to detect static objects and recognise the information they are providing. This paper, in particular, focuses on the computer vision techniques used to detect pedestrians and vehicles. There have been many papers and reviews on pedestrians and vehicles detection so far. However, most of the past papers only reviewed pedestrian or vehicle detection separately. This review aims to present an overview of the AV systems in general, and then review and investigate several detection computer vision techniques for pedestrians and vehicles. The review concludes that both traditional and Deep Learning (DL) techniques have been used for pedestrian and vehicle detection; however, DL techniques have shown the best results. Although good detection results have been achieved for pedestrians and vehicles, the current algorithms still struggle to detect small, occluded, and truncated objects. In addition, there is limited research on how to improve detection performance in difficult light and weather conditions. Most of the algorithms have been tested on well-recognised datasets such as Caltech and KITTI; however, these datasets have their own limitations. Therefore, this paper recommends that future works should be implemented on more new challenging datasets, such as PIE and BDD100K.

show abstract

Section: Techniquesmentioning

confidence: 99%

Pedestrian and Vehicle Detection in Autonomous Vehicle Perception Systems—A Review

Galvao

Abbod

Kalganova

et al. 2021

Sensors

View full text Add to dashboard Cite

show abstract

“…This was done so that the final DNN could be generally applied to different power systems under different conditions (voltage level, source impedance, etc.) [27]. For normalization, every value in the input data was divided by the maximum absolute value and the same procedure was applied to the labels.…”

Section: Pre-processingmentioning

confidence: 99%

A Study on Deep Neural Network-Based DC Offset Removal for Phase Estimation in Power Systems

et al. 2019

View full text Add to dashboard Cite

The purpose of this paper is to remove the exponentially decaying DC offset in fault current waveforms using a deep neural network (DNN), even under harmonics and noise distortion. The DNN is implemented using the TensorFlow library based on Python. Autoencoders are utilized to determine the number of neurons in each hidden layer. Then, the number of hidden layers is experimentally decided by comparing the performance of DNNs with different numbers of hidden layers. Once the optimal DNN size has been determined, intensive training is performed using both the supervised and unsupervised training methodologies. Through various case studies, it was verified that the DNN is immune to harmonics, noise distortion, and variation of the time constant of the DC offset. In addition, it was found that the DNN can be applied to power systems with different voltage levels.

show abstract

“…Dataset dalam penelitian ini, sama dengan yang digunakan oleh penelitian[5]. Arsitektur CNN yang digunakan oleh peneliti adalah 4 conv layer.…”

unclassified

Klasifikasi Ketertarikan Anak PAUD Melalui Ekspresi Wajah Menggunakan Metode CNN

Kusumastuti

Kristian²,

Setyati³

2020

J-TIT

View full text Add to dashboard Cite

The character of emotions in children is different from that of adults, where the characteristics of the emotions in childres include, (1) Briefly and ends suddenly, (2) Seems greater or stronger, (3) Temporary or superficial, (4) Frequent, (5) Can be known clearly from behaviour, and (6) Reaction reflects individuality. Emotions that are felt can be expressed through faces, this is continuous with how interested the child is tho the material presented in fornt of him. Measuring the level of interest in PAUD children in this study using CNN. In the process of training the level of interest in PAUD children, the accuracy value of the four models always increases from epoch 25 until 100 with the highest value being the Rajmehra architecture. But during the data testing process, the architecture in this study increased slightly and the highest peak reached an accuracy value of 81.66%. It is 3.33% better than the result obtained with the Rajmehra architecture. Emosi adalah perasaan atau afeksi yang timbul, ketika seseorang berada dalam suatu keadaan yang dianggap penting oleh individu tersebut. Karakteristik emosi pada anak berbeda dengan karakteristik yang terjadi pada orang dewasa, dimana karekteristik emosi pada anak itu antara lain, (1) Berlangsung singkat dan berakhir tiba-tiba, (2) Terlihat lebih hebat atau kuat, (3) Bersifat sementara atau dangkal, (4) Lebih sering terjadi, (5) Dapat diketahui dengan jelas dari tingkah lakunya, dan (6) Reaksi mencerminkan individualitas. Emosi yang dirasakan dapat diekspresikan melalui wajah, hal ini berkesinambungan dengan seberapa tertariknya anak terhadap tayangan materi yang disajikan dihadapannya. Pengukuran tingkat ketertarikan anak PAUD pada penelitian ini menggunakan CNN. Dalam proses training tingkat ketertarikan anak PAUD, nilai akurasi keempat model selalu mengalami peningkatan mulai dari epoch 25 hingga 100 dengan nilai tertinggi adalah arsitektur Rajmehra. Tetapi saat proses testing data, arsitektur pada penelitian ini mengalami peningkatan secara perlahan dan puncak tertinggi mencapai nilai akurasi 81,66%. Hal tersebut jauh lebih baik 3,33% dibandingkan hasil yang diperoleh dengan arsitektur Rajmehra.

show abstract

Improving the lenet with batch normalization and online hard example mining for digits recognition

Cited by 5 publications

References 4 publications

Pedestrian and Vehicle Detection in Autonomous Vehicle Perception Systems—A Review

Pedestrian and Vehicle Detection in Autonomous Vehicle Perception Systems—A Review

A Study on Deep Neural Network-Based DC Offset Removal for Phase Estimation in Power Systems

Klasifikasi Ketertarikan Anak PAUD Melalui Ekspresi Wajah Menggunakan Metode CNN

Contact Info

Product

Resources

About