CN112487939A - Pure vision light weight sign language recognition system based on deep learning - Google Patents
Pure vision light weight sign language recognition system based on deep learning Download PDFInfo
- Publication number
- CN112487939A CN112487939A CN202011349613.8A CN202011349613A CN112487939A CN 112487939 A CN112487939 A CN 112487939A CN 202011349613 A CN202011349613 A CN 202011349613A CN 112487939 A CN112487939 A CN 112487939A
- Authority
- CN
- China
- Prior art keywords
- sign language
- feature extraction
- layer
- deep learning
- recognition system
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/20—Movements or behaviour, e.g. gesture recognition
- G06V40/28—Recognition of hand or arm movements, e.g. recognition of deaf sign language
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/20—Natural language analysis
- G06F40/205—Parsing
- G06F40/211—Syntactic parsing, e.g. based on context-free grammar [CFG] or unification grammars
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/20—Natural language analysis
- G06F40/279—Recognition of textual entities
- G06F40/284—Lexical analysis, e.g. tokenisation or collocates
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/049—Temporal neural networks, e.g. delay elements, oscillating neurons or pulsed inputs
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/084—Backpropagation, e.g. using gradient descent
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/46—Descriptors for shape, contour or point-related descriptors, e.g. scale invariant feature transform [SIFT] or bags of words [BoW]; Salient regional features
- G06V10/462—Salient features, e.g. scale invariant feature transforms [SIFT]
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Artificial Intelligence (AREA)
- General Health & Medical Sciences (AREA)
- Health & Medical Sciences (AREA)
- General Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Computational Linguistics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Evolutionary Computation (AREA)
- Software Systems (AREA)
- Computing Systems (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Mathematical Physics (AREA)
- Molecular Biology (AREA)
- Multimedia (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Psychiatry (AREA)
- Social Psychology (AREA)
- Human Computer Interaction (AREA)
- Image Analysis (AREA)
Abstract
The invention belongs to the field of sign language recognition, and particularly discloses a pure visual light weight sign language recognition system based on deep learning. The invention can output sentences through operation only by inputting image data without additional information, has simple and efficient overall network structure and short training period, and is suitable for being deployed at a mobile terminal. The program can be operated by only one terminal device, so that the use convenience is greatly improved, and the large-area popularization is facilitated.
Description
Technical Field
The invention relates to the technical field of sign language recognition, in particular to a pure visual light-weight sign language recognition system based on deep learning.
Background
Sign language is an important communication mode between the deaf-mute and the hearing-aid person, and in order to promote the communication convenience between the deaf-mute and the hearing-aid person, it is particularly important to design a sign language recognition system capable of running in real time at a mobile terminal. However, the sign language has rich semantics, and the action amplitude has locality and detail compared with other human behaviors, and is influenced by illumination, background, motion speed and the like, so that the traditional pattern recognition and machine learning method is difficult to realize ideal precision and robustness. In addition, the sign language recognition algorithm in the laboratory environment with a large amount of computation is difficult to deploy in the mobile terminal and realize efficient operation because the hardware conditions of the mobile terminal include a CPU, a GPU, a memory, and the like.
In recent years, image-based deep learning methods have been increasingly successful in continuous phrase sign language recognition tasks. Continuous sentence sign language recognition requires establishing more reliable long-term timing dependencies. Generally, a bidirectional long-time and short-time memory network model is adopted to better model the context semantic information of the long-time sequence of the sign language. Compared with the complexity of the BLSTM network model, the continuous sign language recognition based on the 1-dimensional convolutional network model and the 3-dimensional convolutional network model avoids the complex modeling of the BLSTM network, and saves complex calculated amount on the basis of time sequence modeling. The conventional sign language sentence time sequence segmentation method is complex in process and high in misjudgment rate, and in recent years, scholars gradually bypass time sequence segmentation, introduce a time sequence alignment algorithm CTC in a voice recognition field into the sign language recognition field and achieve good effects.
The existing implementation scheme generally utilizes a bracelet or a glove with a sensor to collect information such as motion and position of a hand, transmits the information to a cloud, extracts hand word information from the information by the cloud through pattern recognition, machine learning or deep learning methods, and finally generates sentences. For the prior art scheme, due to the fact that additional hardware assistance is needed, the method is high in use cost, poor in use convenience, difficult to popularize in a large area, low in identification precision and poor in robustness. The recognition algorithm is complex and is difficult to be deployed on a mobile phone to run in real time.
Aiming at the problems, the invention provides a pure vision light weight sign language recognition system based on deep learning.
Disclosure of Invention
The invention aims to solve the defects in the prior art and provides a pure visual light sign language recognition system based on deep learning.
In order to achieve the purpose, the invention adopts the following technical scheme:
a pure visual light sign language recognition system based on deep learning comprises data acquisition, gesture feature extraction, time sequence feature extraction and sentence generation, wherein the data acquisition is used for acquiring a sign language video to be recognized and preprocessing images, the gesture feature extraction is used for acquiring gesture feature vectors from frames in the sign language video, the time sequence feature extraction is used for extracting sign language word information from a gesture feature vector sequence, and the sentence generation is used for combining all sign language word information into a text sentence according to context;
the identification system further comprises the following use steps:
s1, the application program opens a mobile phone camera to shoot and acquire the sign language video, or directly acquires the sign language video from a folder, and after clicking a start recognition button for a moment, the sign language recognition result is displayed on a screen;
s2, after a sign language video is obtained, firstly, an image sequence obtained by four-time down-sampling is used as source input of a sign language recognition model, an image sequence obtained by eight-time down-sampling is used as source input of a human body detection model, human body coordinates are predicted, then, a source input image is cut by taking a human body as a center and is zoomed to 224 pixels with high width and 224 pixels with high width, and finally, normalization is carried out, and data preparation is finished;
s3, in the gesture feature extraction part, firstly, the first feature extraction layer adopts a 2D convolution layer and a maximum pooling layer for zooming the image, which is beneficial to reducing the calculation amount, and the specific parameters are as follows: the convolution kernel size is 7x7, the step size is 2, the all-zero padding is 3, the channel is 64, the second feature extraction layer adopts two basic residual blocks, and the specific parameters are as follows: convolution kernel size 3x3, step 1, all-zero padding 1, channel 64, the third feature extraction layer uses two basic residual blocks, the specific parameters are: convolution kernel size 3x3, step 1, all-zero padding 1, channel 128, and the fourth feature extraction layer adopts two basic residual blocks, and the specific parameters are: the convolution kernel size is 3x3, the step size is 1, the all-zero padding is 1, the channel is 256, the fifth feature extraction layer adopts two basic residual blocks, and the specific parameters are as follows: convolution kernel size 3x3, step 1, all zero padding 1, channel 512. Finally, a global average pooling layer is followed, and the gesture feature extraction part finally outputs a series of feature vectors with the length of 512;
s4, in the time sequence feature extraction part, firstly accessing a 1D convolutional layer, then accessing a maximum pooling layer, and finally accessing a 1D convolutional layer;
s5, in the sentence generating part, a BLSTM layer is adopted, the obtained sign language word information is used as input, sign language sentence information is output according to the context environment, the sign language sentence information is mapped to a prediction space through a full connection layer, and finally a prediction result can be obtained through CTC beam search decoding.
In another aspect, the present invention provides a mobile phone including the above-mentioned pure visual light sign language recognition system based on deep learning.
In another aspect, the invention provides a tablet computer, which includes the above-mentioned pure visual light sign language recognition system based on deep learning.
In another aspect, the present invention provides a PC computer including the above-mentioned pure visual light sign language recognition system based on deep learning.
In still another aspect, the present invention provides a server including the above-mentioned deep learning-based pure visual lightweight sign language recognition system.
Compared with the prior art, the invention has the beneficial effects that:
the invention ingeniously uses two 1DCNN layers as a short-distance time sequence extractor, thereby outputting sign language word information, and the invention has small operand and high operation speed.
The invention uses a BLSTM layer as a long-distance time sequence extractor, which can capture forward information and backward information, so that the output sentence information is more accurate and smooth. A full connection layer is accessed behind the BLSTM layer, a prediction result is directly output, and simplicity and high efficiency are achieved.
The invention can output sentences through operation only by inputting image data without additional information, has simple and efficient overall network structure and short training period, and is suitable for being deployed at a mobile terminal. The program can be operated by only one terminal device, so that the use convenience is greatly improved, and the large-area popularization is facilitated.
Drawings
Fig. 1 is a framework diagram of an application program in which a sign language recognition system is deployed at a mobile phone end in one embodiment.
FIG. 2 is a flow diagram of a sign language recognition system in one embodiment.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all embodiments.
Examples
Referring to fig. 1-2, the pure visual light weight sign language recognition system based on deep learning provided by the invention comprises data acquisition, gesture feature extraction, time sequence feature extraction and sentence generation, wherein the data acquisition is the acquisition of a sign language video to be recognized and the image preprocessing, the gesture feature extraction is the acquisition of gesture feature vectors from each frame in the sign language video, the time sequence feature extraction is the extraction of sign language word information from a gesture feature vector sequence, and the sentence generation is the combination of all sign language word information into a text sentence according to the context;
the identification system further comprises the following use steps:
s1, the application program opens a mobile phone camera to shoot and acquire the sign language video, or directly acquires the sign language video from a folder, and after clicking a start recognition button for a moment, the sign language recognition result is displayed on a screen;
s2, after a sign language video is obtained, firstly, an image sequence obtained by four-time down-sampling is used as source input of a sign language recognition model, an image sequence obtained by eight-time down-sampling is used as source input of a human body detection model, human body coordinates are predicted, then, a source input image is cut by taking a human body as a center and is zoomed to 224 pixels with high width and 224 pixels with high width, and finally, normalization is carried out, and data preparation is finished;
s3, in the gesture feature extraction part, firstly, the first feature extraction layer adopts a 2D convolution layer and a maximum pooling layer for zooming the image, which is beneficial to reducing the calculation amount, and the specific parameters are as follows: the convolution kernel size is 7x7, the step size is 2, the all-zero padding is 3, the channel is 64, the second feature extraction layer adopts two basic residual blocks, and the specific parameters are as follows: convolution kernel size 3x3, step 1, all-zero padding 1, channel 64, the third feature extraction layer uses two basic residual blocks, the specific parameters are: convolution kernel size 3x3, step 1, all-zero padding 1, channel 128, and the fourth feature extraction layer adopts two basic residual blocks, and the specific parameters are: the convolution kernel size is 3x3, the step size is 1, the all-zero padding is 1, the channel is 256, the fifth feature extraction layer adopts two basic residual blocks, and the specific parameters are as follows: convolution kernel size 3x3, step 1, all zero padding 1, channel 512. Finally, a global average pooling layer is followed, and the gesture feature extraction part finally outputs a series of feature vectors with the length of 512;
s4, in the time sequence feature extraction part, firstly accessing a 1D convolution layer, then accessing a maximum pooling layer, and finally accessing a 1D convolution layer, wherein due to the effects of two layers of convolution and one layer of pooling, the 1DCNN is helpful for extracting short-distance time sequence features, so that sign language word information is output after passing through the 1DCNN layer;
s5, in the sentence generating part, a BLSTM layer is adopted, the obtained sign language word information is used as input, sign language sentence information is output according to the context environment, the sign language sentence information is mapped to a prediction space through a full connection layer, and finally a prediction result can be obtained through CTC beam search decoding.
In another aspect, the present invention provides a mobile phone including the above-mentioned pure visual light sign language recognition system based on deep learning.
In another aspect, the invention provides a tablet computer, which includes the above-mentioned pure visual light sign language recognition system based on deep learning.
In another aspect, the present invention provides a PC computer including the above-mentioned pure visual light sign language recognition system based on deep learning.
In still another aspect, the present invention provides a server including the above-mentioned deep learning-based pure visual lightweight sign language recognition system.
The invention ingeniously uses two 1DCNN layers as a short-distance time sequence extractor, thereby outputting sign language word information, and the invention has small operand and high operation speed. The invention uses a BLSTM layer as a long-distance time sequence extractor, which can capture forward information and backward information, so that the output sentence information is more accurate and smooth. A full connection layer is accessed behind the BLSTM layer, a prediction result is directly output, and simplicity and high efficiency are achieved. The invention adopts the network structure of 2DCNN +1DCNN + BLSTM + CTC, and sentences can be output through operation only by inputting image data without additional information. The overall network structure is simple and efficient, the training period is short, and the method is suitable for being deployed at a mobile terminal. The program can be operated by only one terminal device with a camera, such as a mobile phone, so that the use convenience is greatly improved, and the method is favorable for large-area popularization.
The above description is only for the preferred embodiment of the present invention, but the scope of the present invention is not limited thereto, and any person skilled in the art should be considered to be within the technical scope of the present invention, and the technical solutions and the inventive concepts thereof according to the present invention should be equivalent or changed within the scope of the present invention.
Claims (5)
1. A pure visual light weight sign language recognition system based on deep learning is characterized by comprising data acquisition, gesture feature extraction, time sequence feature extraction and sentence generation, wherein the data acquisition is used for acquiring a sign language video to be recognized and preprocessing images, the gesture feature extraction is used for acquiring gesture feature vectors from frames in the sign language video, the time sequence feature extraction is used for extracting sign language word information from a gesture feature vector sequence, and the sentence generation is used for combining all sign language word information into a text sentence according to context;
the identification system further comprises the following use steps:
s1, the application program opens a mobile phone camera to shoot and acquire the sign language video, or directly acquires the sign language video from a folder, and after clicking a start recognition button for a moment, the sign language recognition result is displayed on a screen;
s2, after a sign language video is obtained, firstly, an image sequence obtained by four-time down-sampling is used as source input of a sign language recognition model, an image sequence obtained by eight-time down-sampling is used as source input of a human body detection model, human body coordinates are predicted, then, a source input image is cut by taking a human body as a center and is zoomed to 224 pixels with high width and 224 pixels with high width, and finally, normalization is carried out, and data preparation is finished;
s3, in the gesture feature extraction part, firstly, the first feature extraction layer adopts a 2D convolution layer and a maximum pooling layer for zooming the image, which is beneficial to reducing the calculation amount, and the specific parameters are as follows: the convolution kernel size is 7x7, the step size is 2, the all-zero padding is 3, the channel is 64, the second feature extraction layer adopts two basic residual blocks, and the specific parameters are as follows: convolution kernel size 3x3, step 1, all-zero padding 1, channel 64, the third feature extraction layer uses two basic residual blocks, the specific parameters are: convolution kernel size 3x3, step 1, all-zero padding 1, channel 128, and the fourth feature extraction layer adopts two basic residual blocks, and the specific parameters are: the convolution kernel size is 3x3, the step size is 1, the all-zero padding is 1, the channel is 256, the fifth feature extraction layer adopts two basic residual blocks, and the specific parameters are as follows: convolution kernel size 3x3, step 1, all zero padding 1, channel 512. Finally, a global average pooling layer is followed, and the gesture feature extraction part finally outputs a series of feature vectors with the length of 512;
s4, in the time sequence feature extraction part, firstly accessing a 1D convolutional layer, then accessing a maximum pooling layer, and finally accessing a 1D convolutional layer;
s5, in the sentence generating part, a BLSTM layer is adopted, the obtained sign language word information is used as input, sign language sentence information is output according to the context environment, the sign language sentence information is mapped to a prediction space through a full connection layer, and finally a prediction result can be obtained through CTC beam search decoding.
2. A handset comprising the deep learning based pure visual lightweight sign language recognition system of claim 1.
3. A tablet computer comprising the deep learning based pure visual lightweight sign language recognition system of claim 1.
4. A PC computer comprising the deep learning based pure visual lightweight sign language recognition system of claim 1.
5. A server comprising the deep learning based pure visual lightweight sign language recognition system of claim 1.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011349613.8A CN112487939A (en) | 2020-11-26 | 2020-11-26 | Pure vision light weight sign language recognition system based on deep learning |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011349613.8A CN112487939A (en) | 2020-11-26 | 2020-11-26 | Pure vision light weight sign language recognition system based on deep learning |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112487939A true CN112487939A (en) | 2021-03-12 |
Family
ID=74935274
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011349613.8A Pending CN112487939A (en) | 2020-11-26 | 2020-11-26 | Pure vision light weight sign language recognition system based on deep learning |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112487939A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113851029A (en) * | 2021-07-30 | 2021-12-28 | 阿里巴巴达摩院(杭州)科技有限公司 | Barrier-free communication method and device |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110348420A (en) * | 2019-07-18 | 2019-10-18 | 腾讯科技（深圳）有限公司 | Sign Language Recognition Method, device, computer readable storage medium and computer equipment |
| CN110458337A (en) * | 2019-07-23 | 2019-11-15 | 内蒙古工业大学 | A kind of net based on C-GRU about vehicle supply and demand prediction method |
| CN111144269A (en) * | 2019-12-23 | 2020-05-12 | 威海北洋电气集团股份有限公司 | Signal-related behavior identification method and system based on deep learning |
| CN111259982A (en) * | 2020-02-13 | 2020-06-09 | 苏州大学 | Premature infant retina image classification method and device based on attention mechanism |
| CN111797777A (en) * | 2020-07-07 | 2020-10-20 | 南京大学 | Sign language recognition system and method based on space-time semantic features |
-
2020
- 2020-11-26 CN CN202011349613.8A patent/CN112487939A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110348420A (en) * | 2019-07-18 | 2019-10-18 | 腾讯科技（深圳）有限公司 | Sign Language Recognition Method, device, computer readable storage medium and computer equipment |
| CN110458337A (en) * | 2019-07-23 | 2019-11-15 | 内蒙古工业大学 | A kind of net based on C-GRU about vehicle supply and demand prediction method |
| CN111144269A (en) * | 2019-12-23 | 2020-05-12 | 威海北洋电气集团股份有限公司 | Signal-related behavior identification method and system based on deep learning |
| CN111259982A (en) * | 2020-02-13 | 2020-06-09 | 苏州大学 | Premature infant retina image classification method and device based on attention mechanism |
| CN111797777A (en) * | 2020-07-07 | 2020-10-20 | 南京大学 | Sign language recognition system and method based on space-time semantic features |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113851029A (en) * | 2021-07-30 | 2021-12-28 | 阿里巴巴达摩院(杭州)科技有限公司 | Barrier-free communication method and device |
| CN113851029B (en) * | 2021-07-30 | 2023-09-05 | 阿里巴巴达摩院(杭州)科技有限公司 | Barrier-free communication method and device |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US10846522B2 (en) | Speaking classification using audio-visual data | |
| Gao et al. | Sign language recognition based on HMM/ANN/DP | |
| Mekala et al. | Real-time sign language recognition based on neural network architecture | |
| CN110070065A (en) | The sign language systems and the means of communication of view-based access control model and speech-sound intelligent | |
| EP3876140A1 (en) | Method and apparatus for recognizing postures of multiple persons, electronic device, and storage medium | |
| CN110348420A (en) | Sign Language Recognition Method, device, computer readable storage medium and computer equipment | |
| CN113835522A (en) | Sign language video generation, translation and customer service method, device and readable medium | |
| CN112001248B (en) | Active interaction method, device, electronic equipment and readable storage medium | |
| KR20120120858A (en) | Service and method for video call, server and terminal thereof | |
| CN110992783A (en) | Sign language translation method and translation equipment based on machine learning | |
| Balasuriya et al. | Learning platform for visually impaired children through artificial intelligence and computer vision | |
| CN113723327A (en) | Real-time Chinese sign language recognition interactive system based on deep learning | |
| Wang et al. | (2+ 1) D-SLR: an efficient network for video sign language recognition | |
| CN112487939A (en) | Pure vision light weight sign language recognition system based on deep learning | |
| KR102377767B1 (en) | Handwriting and arm movement learning-based sign language translation system and method | |
| CN112487951B (en) | Sign language recognition and translation method | |
| CN111368800A (en) | Gesture recognition method and device | |
| CN108628454B (en) | Visual interaction method and system based on virtual human | |
| CN113420783B (en) | Intelligent man-machine interaction method and device based on image-text matching | |
| JP2021114313A (en) | Face composite image detecting method, face composite image detector, electronic apparatus, storage medium and computer program | |
| Wang et al. | A bi-directional interactive system of sign language and visual speech based on portable devices | |
| CN117456063B (en) | Face driving method and device based on voice, electronic equipment and storage medium | |
| Rajendrababu et al. | Design and implementation of smart book reader for the blind | |
| CN114356076B (en) | Gesture control method and system | |
| Perera et al. | Finger spelled Sign Language Translator for Deaf and Speech Impaired People in Srilanka using Convolutional Neural Network |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20210312 |
|
| RJ01 | Rejection of invention patent application after publication |