The heavy burdens of computation and off-chip traffic impede deploying the large scale convolution neural network on embedded platforms. As CNN is attributed to the strong endurance to computation errors, employing block floating point (BFP) arithmetics in CNN accelerators could save the hardware cost and data traffics efficiently, while maintaining the classification accuracy. In this paper, w...

• We use three popular CNN models: VGG16, VGG19 and AlexNet. • To limit rescaling, the last pooling stage is removed and the stride of the last but one pooling layer is decreased. • We take feature maps at three different locations of the FCN, and concatenate them to form a single tensor. • A 3 x 3 convolutional layer learns 64 saliency-specific feature maps, then a 1 x 1 convolution learns to ...

In the x-ray diffraction experiment, we collected about 8 million diffraction images in total, where less then 0.3% (20 thousand) were “good” images that could be used for further scientific study. Even among those “good” images, about 95% was “non-diffuse” image (1b, 1d, 1f), and only 5% (1 thousand) was “diffuse” image (1a, 1c, 1e) that we decided to extract out using machine learning methods...

Pisang cavendish banyak dikonsumsi di Indonesia dan berpotensi menjadi komoditas utama Indonesia. Namun, proses pemilihan kualitas pisang masih yang dilakukan secara tradisional. Hal ini penghambat dalam utama. Klasifikasi mutu modern dapat untuk memperbaiki seleksi meningkatkan penjualan sektor pertanian. Peningkatan sector pertanian akan menjadikan sebagai ekonomi Metode deep learning yaitu C...

Herbal leaves are a type that is often used by people in the health sector. The problem faced lack of knowledge about types herbal and difficulty distinguishing for ordinary who do not understand plants. If any plant used, it will have negative impact on health. Automatic classification with help technology reduce risk misidentification leaf types. To make identification, precise accurate detec...

Many techniques related to image analysis have been proposed by researchers which are being used detect a large number of diseases. These images carefully analyzed radiologists and doctors, after careful interpretation, the results obtained finally help in making an appropriate diagnosis. This is complicated time consuming task, requires high levels concentration. Therefore, experts who analyze...

This paper deals with the classification of kidneys for renal stones on ultrasound images. Convolutional neural network (CNN) and pre-trained CNN (VGG16) models are used to extract features from Extreme gradient boosting (XGBoost) classifiers random forests classification. The extracted VGG16 compare performance XGBoost forest. An image normal was classified. work uses 630 real images Al-Diwani...

Deep convolutional neural network (CNN) inference requires significant amount of memory and computation, which limits its deployment on embedded devices. To alleviate these problems to some extent, prior research utilize low precision fixed-point numbers to represent the CNN weights and activations. However, the minimum required data precision of fixed-point weights varies across different netw...

In recent years, brain tumors become the leading cause of death in the world. Detection and rapid classification of this tumor are very important and may indicate the likely diagnosis and treatment strategy. In this paper, we propose deep learning techniques based on the combinations of pre-trained VGG-16 CNNs to classify three types of brain tumors (i.e., meningioma, glioma, and pituitary tumo...