FazBrowse GitHub Viewer | Trending |
URL:
| Home
Tools: [Download Repo ZIP]   [Original HTTPS Page]

Optimize in time the image feature extractions algorithms by hakim-cherif-mchp · Pull Request #23 · edgeimpulse/linux-sdk-python · GitHub

/ linux-sdk-python Public

Optimize in time the image feature extractions algorithms #23

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
hakim-cherif-mchp wants to merge 3 commits into
base: master
Choose a base branch
Could not load branches
Branch not found: {{ refName }}

Are you sure you want to change the base?

Some commits from the old base branch may be removed from the timeline, and old review comments may become outdated.
from
Conversation 6 Commits 3 Checks 0 Files changed
Open

Optimize in time the image feature extractions algorithms #23

Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
2ae613f
Update image resize interpolation from INTER_AREA to INTER_NEAREST as…
hakim-cherif-mchp Sep 6, 2024
9fc4ef4
Using Numpy arrays instead of lists as this is significantly quicker …
hakim-cherif-mchp Sep 6, 2024
e676286
Correct an error from if self.isGrayscale: to if is_grayscale:
hakim-cherif-mchp Oct 25, 2024
File filter

Filter by extension

Filter by extension .py  (1) All 1 file type selected
Viewed files
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Unified
Split
Hide whitespace
Diff view
Unified
Split
Hide whitespace
47 changes: 24 additions & 23 deletions edge_impulse_linux/image.py
Show comments
View file
Open in desktop
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters. Learn more about bidirectional Unicode characters
Show hidden characters
Original file line number Diff line number Diff line change
Expand Up @@ -92,7 +92,7 @@ def get_features_from_image(self, img, crop_direction_x='center', crop_direction
# One dim will match the classifier size, the other will be larger
resize_size = (resize_size_w, resize_size_h)

resized = cv2.resize(img, resize_size, interpolation=cv2.INTER_AREA)
resized = cv2.resize(img, resize_size, interpolation=cv2.INTER_NEAREST)

Alex-EEE Oct 24, 2024

Copy link
Copy Markdown
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Choose a reason Spam Abuse Off Topic Outdated Duplicate Resolved Low Quality

We feel Inter Area is the best trade off for performance and output, especially since the user should only be downsizing. But I'm open to hear your thoughts

hakim-cherif-mchp Dec 24, 2024

Copy link
Copy Markdown
Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Choose a reason Spam Abuse Off Topic Outdated Duplicate Resolved Low Quality

Hi Alex,
Sorry for the delay, I tested more extensively the two algorithms, and it appears that it is possible to get a very high similarity score between the two sets of extracted features, but only under certain circumstances.
So, in my opinion, you should keep the INTER_AREA.

if (crop_direction_x == 'center'):
crop_x = int((resize_size_w - EI_CLASSIFIER_INPUT_WIDTH) / 2) # 0 when same
Expand All @@ -117,19 +117,20 @@ def get_features_from_image(self, img, crop_direction_x='center', crop_direction

if self.isGrayscale:
cropped = cv2.cvtColor(cropped, cv2.COLOR_BGR2GRAY)
pixels = np.array(cropped).flatten().tolist()
pixels = np.array(cropped).flatten().astype(np.uint32)

features.extend((pixels << 16) + (pixels << 8) + pixels)

for p in pixels:
features.append((p << 16) + (p << 8) + p)
else:
pixels = np.array(cropped).flatten().tolist()
pixels = np.array(cropped).reshape(-1, 3)

for ix in range(0, len(pixels), 3):
r = pixels[ix + 0]
g = pixels[ix + 1]
b = pixels[ix + 2]
features.append((r << 16) + (g << 8) + b)
r = pixels[:, 0].astype(np.uint32)
g = pixels[:, 1].astype(np.uint32)
b = pixels[:, 2].astype(np.uint32)

features = (r << 16) + (g << 8) + b

features=features.tolist()
return features, cropped

def get_features_from_image_auto_studio_setings(self, img):
Expand Down Expand Up @@ -211,27 +212,27 @@ def get_features_from_image_with_studio_mode(img, mode, output_width, output_hei
offset = (in_frame_rows - new_height) // 2
cropped_img = img[offset:offset + new_height, :]

resized_img = cv2.resize(cropped_img, (output_width, output_height), interpolation=cv2.INTER_AREA)
resized_img = cv2.resize(cropped_img, (output_width, output_height), interpolation=cv2.INTER_NEAREST)
elif mode == 'fit-longest':
resized_img = resize_with_letterbox(img, output_width, output_height)
elif mode == 'squash':
resized_img = cv2.resize(img, (output_width, output_height), interpolation=cv2.INTER_AREA)
resized_img = cv2.resize(img, (output_width, output_height), interpolation=cv2.INTER_NEAREST)
else:
raise ValueError(f"Unsupported mode: {mode}")

if is_grayscale:
resized_img = cv2.cvtColor(resized_img, cv2.COLOR_BGR2GRAY)
pixels = np.array(resized_img).flatten().tolist()
cropped = cv2.cvtColor(cropped, cv2.COLOR_BGR2GRAY)
pixels = np.array(cropped).flatten().astype(np.uint32)

features.extend((pixels << 16) + (pixels << 8) + pixels)
else:
pixels = np.array(resized_img).reshape(-1, 3)

for p in pixels:
features.append((p << 16) + (p << 8) + p)
else:
pixels = np.array(resized_img).flatten().tolist()
r = pixels[:, 0].astype(np.uint32)
g = pixels[:, 1].astype(np.uint32)
b = pixels[:, 2].astype(np.uint32)

for ix in range(0, len(pixels), 3):
r = pixels[ix + 0]
g = pixels[ix + 1]
b = pixels[ix + 2]
features.append((r << 16) + (g << 8) + b)
features = (r << 16) + (g << 8) + b

features=features.tolist()
return features, resized_img
Back | FazBrowse Home | New Git URL