diff options
Diffstat (limited to 'site/public/research/01_from_1_to_100_pixels')
| -rw-r--r-- | site/public/research/01_from_1_to_100_pixels/index.html | 4 |
1 files changed, 4 insertions, 0 deletions
diff --git a/site/public/research/01_from_1_to_100_pixels/index.html b/site/public/research/01_from_1_to_100_pixels/index.html index 7b86f5ef..cc9d3f94 100644 --- a/site/public/research/01_from_1_to_100_pixels/index.html +++ b/site/public/research/01_from_1_to_100_pixels/index.html @@ -92,6 +92,10 @@ <li>100x100 all you need for medical diagnosis</li> <li>100x100 0.5% of one Instagram photo</li> </ul> +<p>Notes:</p> +<ul> +<li>Google FaceNet used images with (face?) sizes: Input sizes range from 96x96 pixels to 224x224pixels in our experiments. FaceNet: A Unified Embedding for Face Recognition and Clustering <a href="https://arxiv.org/pdf/1503.03832.pdf">https://arxiv.org/pdf/1503.03832.pdf</a></li> +</ul> <p>Ideas:</p> <ul> <li>Find specific cases of facial resolution being used in legal cases, forensic investigations, or military footage</li> |