Skip to content
Open
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
62 changes: 62 additions & 0 deletions OUPUT_EXPLANATION.md
Original file line number Diff line number Diff line change
@@ -0,0 +1,62 @@
# FashionFormer Output Explanation
The result data structure is organized as a tuple containing three lists.
Here is a sample image for which the model output is generated.

![sample image](figs/sample_image.png?raw=true)


## First Item: Garment Detection Information
Data Type: List of Arrays
Length: 46
Array Dimensionality: (X, 5)
X: Number of times the particular garment was detected within the image

Each row in an array represents a detected instance of the garment.
The first 4 numbers in each row are always 0 (potentially reserved for bounding box coordinates?).
The last number in each row indicates the certainty level of the garment detection.
If the garment was not detected (probability below a certain threshold), the corresponding array's dimensionality is (0, 5).

```
result[0][4]
array([[0. , 0. , 0. , 0. , 0.76458955],
[0. , 0. , 0. , 0. , 0.12469797],
[0. , 0. , 0. , 0. , 0.10306859]],
dtype=float32)
```

## Second Item: Pixel Mask for Garment Detection
Data Type: List of Boolean Arrays
Length: 46
Array Dimensionality: same dimensionality as input image
Array Length: X (number of detected instances of garments in the image)

Each Boolean array has the same dimensionality as the input image.
True values correspond to pixels where the corresponding garment was detected.

```
result[1][4][0]
# Note how this image corresponds to result[0][4][0]
```
![Alt text](figs/sample_mask_1.png?raw=true)

```
result[1][4][1]
# Note how this Image corresponds to result[0][4][1]
```
![Alt text](figs/sample_mask_2.png?raw=true)

```
result[1][4][2]
# Note how this Image corresponds to result[0][4][2]
```
![Alt text](figs/sample_mask_3.png?raw=true)


## Third Item: Attribute Detection Information
Data Type: List of Arrays
Length: 46
Array Dimensionality: (294,)

Each array contains probabilities of 294 fashion attributes being present in the detected garment.
Each of the 294 numbers represents the probability of a specific fashion attribute being present in the garment.

Binary file added figs/sample_image.png
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
Binary file added figs/sample_mask_1.png
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
Binary file added figs/sample_mask_2.png
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
Binary file added figs/sample_mask_3.png
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.