diff --git a/README.md b/README.md
index 3ed75f2..4dc827c 100644
--- a/README.md
+++ b/README.md
@@ -244,7 +244,7 @@ face_objs = DeepFace.extract_faces(img_path = "img.jpg",
Face recognition models are actually CNN models and they expect standard sized inputs. So, resizing is required before representation. To avoid deformation, deepface adds black padding pixels according to the target size argument after detection and alignment.
-
+
[RetinaFace](https://sefiks.com/2021/04/27/deep-face-detection-with-retinaface-in-python/) and [MTCNN](https://sefiks.com/2020/09/09/deep-face-detection-with-mtcnn-in-python/) seem to overperform in detection and alignment stages but they are much slower. If the speed of your pipeline is more important, then you should use opencv or ssd. On the other hand, if you consider the accuracy, then you should use retinaface or mtcnn.
diff --git a/icon/detector-outputs-20230203.jpg b/icon/detector-outputs-20230203.jpg
new file mode 100644
index 0000000..b894e75
Binary files /dev/null and b/icon/detector-outputs-20230203.jpg differ
diff --git a/icon/detector-portfolio-v3.jpg b/icon/detector-portfolio-v3.jpg
deleted file mode 100644
index d67642a..0000000
Binary files a/icon/detector-portfolio-v3.jpg and /dev/null differ
diff --git a/icon/look-alike-v3.jpg b/icon/look-alike-v3.jpg
deleted file mode 100644
index 19d439c..0000000
Binary files a/icon/look-alike-v3.jpg and /dev/null differ
diff --git a/icon/parental-look-alike-v2.jpg b/icon/parental-look-alike-v2.jpg
deleted file mode 100644
index bd08ddf..0000000
Binary files a/icon/parental-look-alike-v2.jpg and /dev/null differ
diff --git a/icon/tech-stack-v2.jpg b/icon/tech-stack-v2.jpg
deleted file mode 100644
index 32a69f7..0000000
Binary files a/icon/tech-stack-v2.jpg and /dev/null differ