krishnapal2308 commited on
Commit
599f168
1 Parent(s): 0d65c8f

Add Description

Browse files
app.py CHANGED
@@ -8,12 +8,28 @@ import warnings
8
 
9
  warnings.filterwarnings('ignore')
10
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
11
 
12
  def process_image_and_generate_output(image, model_selection):
13
  if image is None:
14
  return "Please select an image", None
15
- # (Trained only for 15 epochs without any hyperparameter tuning, utilizing inception v3)'
16
- # (SOTA model for Image captioning)
17
  if model_selection == 'Basic Model':
18
  result = inference_script.evaluate(image)
19
  pred_caption = ' '.join(result).rsplit(' ', 1)[0]
@@ -46,7 +62,6 @@ sample_images = [
46
  [os.path.join(os.path.dirname(__file__), "sample_images/3.jpg"), 'Basic Model']
47
  ]
48
 
49
-
50
  # Create a dropdown to select sample image
51
  image_input = gr.Image(label="Upload Image")
52
 
@@ -61,7 +76,7 @@ iface = gr.Interface(fn=process_image_and_generate_output,
61
  examples=sample_images,
62
  cache_examples=True,
63
  allow_flagging='never',
64
- title="Eye For Blind | Image Captioning & TTS",
65
- description="To be added")
66
 
67
  iface.launch()
 
8
 
9
  warnings.filterwarnings('ignore')
10
 
11
+ # Define problem statement
12
+ problem_statement = """
13
+ ### Problem Statement
14
+ Visually impaired individuals face challenges in understanding image content. This project aims to address this issue by generating descriptive spoken captions for images, leveraging CNNs and RNNs for feature extraction and sequence generation, respectively. The model is trained on the Flickr8K dataset and extended with an attention mechanism for enhanced accessibility.
15
+ """
16
+
17
+ # Define solution overview
18
+ solution_overview = """
19
+ ### Solution Overview
20
+ The basic model, trained for a limited duration without extensive hyperparameter tuning, primarily focuses on exploring subclassing techniques. To improve inference quality, Vit-GPT2 architecture is integrated. [Visit the Kaggle notebook](https://www.kaggle.com/code/krishna2308/eye-for-blind) for implementation details.
21
+ """
22
+
23
+ # Define real-life scenario application
24
+ real_life_scenario = """
25
+ ### Real-life Scenario Application
26
+ While this current implementation may not support real-time processing, the potential for future development is vast. Where a visually impaired individual wears smart glasses equipped with a camera. As they move around, the camera captures live footage of their surroundings, which is then processed in real-time by the image captioning model integrated into the glasses. The generated spoken descriptions can be streamed directly to the user's earpiece, providing instant audio feedback about their environment.
27
+ """
28
+
29
 
30
  def process_image_and_generate_output(image, model_selection):
31
  if image is None:
32
  return "Please select an image", None
 
 
33
  if model_selection == 'Basic Model':
34
  result = inference_script.evaluate(image)
35
  pred_caption = ' '.join(result).rsplit(' ', 1)[0]
 
62
  [os.path.join(os.path.dirname(__file__), "sample_images/3.jpg"), 'Basic Model']
63
  ]
64
 
 
65
  # Create a dropdown to select sample image
66
  image_input = gr.Image(label="Upload Image")
67
 
 
76
  examples=sample_images,
77
  cache_examples=True,
78
  allow_flagging='never',
79
+ title="Eye For Blind | Image Captioning & TTS Demo",
80
+ description=f"{problem_statement}\n\n{solution_overview}\n\n{real_life_scenario}")
81
 
82
  iface.launch()
gradio_cached_examples/17/log.csv ADDED
@@ -0,0 +1,5 @@
 
 
 
 
 
 
1
+ output 0,output 1,flag,username,timestamp
2
+ people sitting on top of a snow covered slope,"{""path"":""gradio_cached_examples\\17\\output 1\\ca5aaabdfeef6bfd1a0b\\audio"",""url"":null,""size"":null,""orig_name"":""audio"",""mime_type"":null}",,,2024-02-09 15:43:56.019598
3
+ five children getting ready to sled,"{""path"":""gradio_cached_examples\\17\\output 1\\616c31df069fc6f8ff8d\\audio"",""url"":null,""size"":null,""orig_name"":""audio"",""mime_type"":null}",,,2024-02-09 15:43:57.007513
4
+ a man riding a bike down a mountain side,"{""path"":""gradio_cached_examples\\17\\output 1\\a7b3fb069783e183f30e\\audio"",""url"":null,""size"":null,""orig_name"":""audio"",""mime_type"":null}",,,2024-02-09 15:43:58.457672
5
+ a man on a bicycle on a mountain,"{""path"":""gradio_cached_examples\\17\\output 1\\5f5ac48c51830807ebc1\\audio"",""url"":null,""size"":null,""orig_name"":""audio"",""mime_type"":null}",,,2024-02-09 15:43:59.021601
gradio_cached_examples/17/output 1/5f5ac48c51830807ebc1/audio ADDED
Binary file (11.2 kB). View file
 
gradio_cached_examples/17/output 1/616c31df069fc6f8ff8d/audio ADDED
Binary file (11 kB). View file
 
gradio_cached_examples/17/output 1/a7b3fb069783e183f30e/audio ADDED
Binary file (13 kB). View file
 
gradio_cached_examples/17/output 1/ca5aaabdfeef6bfd1a0b/audio ADDED
Binary file (13.2 kB). View file