Air Writing Detection & Recognition Using Deep Learning
A non-touch system is a modern approach to computer-interface technology that will revolutionize human-computer interaction. The interface will allow the user to enter data and interact with a human, machine, or robot in an uncontrolled environment, treatment, or industrial life. How
2025-06-28 16:25:04 - Adil Khan
Air Writing Detection & Recognition Using Deep Learning
Project Area of Specialization Artificial IntelligenceProject SummaryA non-touch system is a modern approach to computer-interface technology that will revolutionize human-computer interaction. The interface will allow the user to enter data and interact with a human, machine, or robot in an uncontrolled environment, treatment, or industrial life. However, it is difficult to enter data into the machine and interact with humans and machines with a variety of complexities such as cluttered environment, gesture tracking, and speed.
There are many evolving systems, e.g., aerial handwriting, sign language recognition, and finger alphabet recognition, that will require substantial effort for all character learning and overhead processing, hence the classification accuracy is reduced. Therefore, this project proposes a contactless character writing system that allows users to use a virtual keyboard (tracing hand movements) and save the detected alphabets on the notepad. We divide this work into three steps.
- Detecting the object.
- Tracking object movement from frame to frame.
- Analyzing the behavior of an object.
- Save and display the analyzed word on a notepad.
- To make the educational and communicational system smart according to modern world technologies.
- To introduce a smart way of interacting with the system.
- To optimize and update the performance of typing.
- To reduce the typing errors as writing in the air will reduce typing mistakes.
- To minimize the effort required to input text by patients with severe motor disabilities.
The system proposed in this method will consist of the following steps. It will track the motion of a color object tip, plot the motion of the object tip, optical character organization (OCR) will be applied to the plotted image, and the output will be matched with the trained local database for OCR and the most possible match will be achieved and then displayed on the notepad. This project will be based on python and the libraries used will be OPEN CV and TensorFlow.
Benefits of the Project- To make the educational and communicational system smart according to modern world technologies. e.g. As in educational institutes teachers will be able to interact seamlessly with the system just by writing the desired word in the air instead of typing on a physical keyboard.
- To help the disabled people (fingers disability) to live their life better by providing theme virtual keyboard by simply moving their hand in the air to enter text more easily and more efficiently than with a traditional keyboard as we will put a mark on their wrist, detect it by the system and then draw their desirable alphabet and store it on a notepad.
Hardware:
Jetson Nano Developer Kit :
A system that consists of graphical memory for the training of the dataset.
Specifications:
| GPU | 128-core NVIDIA Maxwell™ |
| CPU | Quad-core ARM® A57 @ 1.43 GHz |
| Memory | 4 GB 64-bit LPDDR4 25.6 GB/s |
| Storage | microSD (Card not included) |
| Video Encode | 4Kp30 | 4x 1080p30 | 9x 720p30 (H.264/H.265) |
| Video Decode | 4Kp60 | 2x 4Kp30 | 8x 1080p30 | 18x 720p30 (H.264/H.265) |
| Connectivity | Gigabit Ethernet, 802.11ac wireless† |
| Camera | 1x MIPI CSI-2 connector |
| Display | HDMI |
| USB | 1x USB 3.0 Type-A,2x USB 2.0 Type-A, USB 2.0 Micro-B |
| Others | 40-pin header (GPIO, I2C, I2S, SPI, UART) |
| Mechanical | 100 mm x 80 mm x 29 mm |
Camera:
Arducam IMX219 Camera Module with fisheye lens for Jetson Nano and Raspberry Pi Compute Module.
Features:
- Sony 8MP IMX219 Sensor
- Optical Format: 1/4 inch
- Frame Rate: 30fps@8MP, 60fps@1080p, 180fps@720p
- Data Format: RAW8/RAW10
- Lens part number:M32076M20
- EFL: 0.76mm
- F.NO: 2.1
- Focus Type: Fixed Focus
- View Angle: 220(H)
- Interface: MIPI CSI-2 2-lane/4-lane
- IR Sensitivity: Integral IR Filter, visible light only
GPU
CPU
Memory
Storage
Video Encode
Video Decode
Connectivity
Camera
Display
USB
Others
Mechanical
Final Deliverable of the Project HW/SW integrated systemCore Industry EducationOther Industries IT , Telecommunication Core Technology Artificial Intelligence(AI)Other Technologies Robotics, Big DataSustainable Development Goals Quality EducationRequired Resources| Elapsed time in (days or weeks or month or quarter) since start of the project | Milestone | Deliverable |
|---|---|---|
| Month 1 | Proposal WritingLiterature Review | 2022-02-12to2022-02-28 |
| Month 2 | Development of Python Program | 2022-03-01to2022-03-30 |
| Month 3 | Design Data set models | 2022-04-01to2022-04-30 |
| Month 4 | Experimentation and testing | 2022-05-01to2022-05-28 |
| Month 5 | Thesis Write up | 2022-06-01to2022-06-29 |
| Month 6 | Submission of Paper and Thesis | 2022-07-10 |