v0.3.2

• fixed predict with certain image resolution (#45)
• added resize_method parameter (#42)
• fixed readme python example referring to a renamed method (#35)
• added optional --source-threaded (#34)
• added --source-fps argument (#33)
• added optional --mask-cache-time (#32)
• various dependency bumps

v0.3.1

• fixed resnet urls (#31)

v0.3.0

• added project goals to readme (#30)
• added TF_CPP_MIN_LOG_LEVEL to cli (#29)
• renamed image-to-mask to draw-mask (#28)
• added blur-background (#27)
• fixed bgr to rgb color conversion (#25)
• added list-models CLI sub command (#24)

v0.2.1

• added args for source image size and fourcc (#23) …
• added optional mask blur (#22) …
• replace background (#21) …
• input source as context manager (#20) …
• fixed mask not shown if not colored (#19)
• added --internal-resolution and --output-stride (#18) …
• reduced camera latency (#17) …
• improve timing logging (#16) …
• added simple virtual webcam support (#15)

v0.1.1

• minimal webcam support

v0.1.0

• added support for body part segmentation
• moved "all imports" to tf_bodypix.api
• moved dependencies to extras

v0.0.4

• pypi release

v0.0.3

• pypi release

v0.0.2

• pypi package

v0.0.1

• initial release