a matrix that moves over the input data, performs the dot product with the sub-region of input data, and gets the output as the matrix of dot products.
Example: Then the number of channels for each convolution kernel and the computation complexity can be decreased. [3]
en: ядро свертки
decoder network
phrase
Maps the feature representation back into the input data space.
Example: Most recent deep architectures for segmentation have identical encoder networks, i.e, VGG16, but differ in the form of the decoder network, training and inference. [2]
en: сеть декодера
dense connection
phrase
a type of layer in a deep neural network that use a linear operation where every input is connected to every output by a weight.
Example: The dense connection between fire modules and the passthrough layer are then properly used in Tinier-YOLO with the purpose to get higher detection accuracy.
en: плотное соединение
Discrete Cosine Transformation (DCT)
phrase
The DCT is a technique allowing the conversion of a signal into elementary frequency components.
Example: In the final step, we have to measure the DCT (Discrete Cosine Transformation) for the energy of the log filter bank. [1]
en: Дискретное косинусное преобразование (DCT)
Evaluation
noun
the making of a judgement about the amount, number, or value of something; assessment.
Example: In this study (see below), we have ported the original annotations given with the dataset used for evaluation into JAMS files and made them accessible on the internet together with the JAMS files after deformation. [1]
en: Оценка
Fast Fourier Transformation (FFT)
phrase
an algorithm that calculates the discrete Fourier transform (DFT) of some sequence – the discrete Fourier transform is a tool to convert specific types of sequences of functions into other types of representations.
Example: Fast Fourier Transformation (FFT) is then used for each frame to evaluate the frame response. [1]
en: Быстрое преобразование Фурье (БПФ)
Feature
noun
a distinctive attribute or aspect of something.
Example: Notice that we have used a small feature map (5, 5) in k1 proportional to the input sizes (256, 256) is made to enable the systems to learn small, scattered signals which can be combined at successive layers to collect arguments to confirm larger ’time–frequency signs’ that indicate the influence of various sound groups, even though acoustic masking occurs. [1]
en: Особенность
feature map
phrase
The feature map is the output of one filter applied to the previous layer.
Example: This is primarily because max pooling and sub-sampling reduce feature map resolution. [2]
en: карта объектов
Fine-tuning
noun
make small adjustments to (something) in order to achieve the best or a desired performance.
Example: We have used a greedy-based pre-training approach with fine-tuning to train the De-noising Auto Encoder. [1]
en: Тонкая настройка
low-level vision cues
phrase
A given visual environment
Example: Early methods that relied on low-level vision cues have fast been superseded by popular machine learning algorithms. [2]
en: зрительные сигналы низкого уровня
max pooling
phrase
a pooling operation that calculates the maximum, or largest, value in each patch of each feature map.
Example: This is primarily because max pooling and sub-sampling reduce feature map resolution. [2]
en: максимальное объединение
network pruning
phrase
practice of removing parameters (which may entail removing individual parameters, or parameters in groups such as by neurons) from an existing network.
Example: Network pruning and sparsity is to prune unimportant neuron connection. [3]
en: обрезка сети
pixel-wise
phrase
one pixel at a time.
Example: Our architecture, SegNet, is designed to be an efficient architecture for pixel-wise semantic segmentation. [2]
en: попиксельный
Semantic segmentation
phrase
Semantic segmentation is a natural step in the progression from coarse to fine inference
Example: Semantic segmentation has a wide array of applications ranging from scene understanding, inferring support-relationships among objects to autonomous driving. [2]
en: Семантическая сегментация
sub-sampling
noun
a method that reduces data size by selecting a subset of the original data.
Example: This is primarily because max pooling and sub-sampling reduce feature map resolution. [2]