Atari Jaguar2

Florence-2 is a versatile Vision-Language Model (VLM), capable of handling multiple vision tasks within a single model. Its zero-shot capabilities are impressive across diverse tasks such as image captioning, object detection, segmentation and OCR. While Florence-2 performs well out-of-the-box, additional fine-tuning can further adapt the model to new tasks or improve its performance on unique, custom datasets.

Our proprietary infrared coatings seal these hygroscopic crystals, protecting them from moisture and extending their operational lifespan while simultaneously enhancing optical performance. These coatings ensure that your infrared optics maintain their performance even in challenging conditions.

Florence-2 can also generate segmentation polygons grounded by text ('') or by bounding boxes (''):

Florence-2 demonstrates strong OCR capabilities. It can extract text from an image with the '' task prompt, and extract both text and its location with '' :

MobileVLM V2

We produce more than just circles and rectangular shaped components. Challenge us to produce a complex shape or pattern demanded by your application.

At Omega Optical, we are committed to advancing the field of FT-IR spectroscopy through innovative optical solutions. Our expertise in optical filters, diffraction gratings, infrared optics, and coatings ensures that you have the tools you need to achieve unparalleled precision and accuracy in your research and analysis.

Despite its relatively small size, with versions of 0.23B & 0.77B parameters, Florence-2 achieves state-of-the-art (SOTA) performance. Its compact size enables efficient deployment on devices with limited computing resources, while ensuring fast inference speeds.

Omega Optical offers a variety of advanced infrared optics from specialty materials which enhance the performance of FT-IR instruments:

Samsung DVD-N2000

Florence-2 can also perform text-grounded object detection. By providing specific object names or descriptions as input, Florence-2 detects bounding boxes around the specified objects.

The model was pre-trained on an enormous, high quality dataset called FLD-5B, consisting of 5.4B annotations on 126 million images. This allows Florence-2 to excel in zero-shot performance on many tasks without requiring additional training.

Florence-2 can identify densely packed regions in the image, and to provide their bounding box coordinates and their related labels or captions. To extract bounding boxes with labels, use the ’’task prompt:

Fourier Transform Infrared (FT-IR) Spectroscopy is a pivotal technique in modern analytical science, enabling precise identification and quantification of materials. At the heart of FT-IR lies the principle of transforming infrared light into detailed spectral data, providing insights into the molecular composition of samples. Omega Optical is at the forefront of this technology, offering cutting-edge solutions that enhance the capabilities of FT-IR instruments.

Neon light synthesizer

Trip a Tron

Florence-2 can generate image captions at various levels of detail, using the '' , '' or '' task prompts.

Omega Optical produces diamond turned off-axis parabolas in a variety of configurations, as well as coated or uncoated. We are specialists in producing OAP's with excellent mechanical or optical datums to ensure easy, precise, and repeatable alignment.

In recent years, the field of computer vision has witnessed the rise of foundation models that enable image annotation without the need for training custom models. We’ve seen models like CLIP [2] for classification, GroundingDINO [3] for object detection, and SAM [4] for segmentation — each excelling in its domain. But what if we had a single model capable of handling all these tasks together?

Learn more about the full suite of capabilities which are trusted by leading Global OEM’s and their Start-up challengers to deliver the right photons, to the right place, at the right time.

In this case, the '' identified the lens, while the '' was less specific. However, this performance may vary with different images.

Florence-2 was released by Microsoft in June 2024. It was designed to perform multiple vision tasks within a single model. It is an open-source model, available on Hugging Face under the permissive MIT licence.

Llamasoft Wiki

Polkadot beamsplitters are used Fourier Transform Infrared (FT-IR) Spectroscopy, where they function by dividing the incident infrared light into two paths—one directed towards the sample and the other towards a reference. This division allows for simultaneous measurement of the sample and reference signals, enhancing the accuracy and efficiency of the spectral data collection. The polkadot pattern, with its precise distribution of reflective and transmissive regions, ensures consistent and uniform splitting of the infrared light. This uniformity is crucial for maintaining the integrity of the interferometric measurements, leading to highly accurate and reliable FT-IR spectroscopic analysis.

Florence-2 can generate captions for specific regions of an image defined by bounding boxes. For this, it takes the bounding box location as input. You can extract the category with '' or a description with '' .

Our proprietary infrared coatings are engineered to provide optimal protection for infrared optics while maintaining their performance. These coatings are specifically designed to seal hygroscopic crystals, such as KBr and CsI, against environmental factors. By preventing moisture ingress, our coatings extend the longevity and reliability of your FT-IR components.

The run_example function combines the task prompt with any additional text input (if provided) into a single prompt. Using the processor, it generates text and image embeddings that serve as inputs to the model. The magic happens during the model.generate step, where the model’s response is generated. Here’s a breakdown of some key parameters:

Goom audio visualization

The model accurately describes the image and its surrounding. It even identifies the camera’s brand and model, demonstrating its OCR ability. However, in the '' task there are minor inconsistencies, which is expected from a zero-shot model.

In this tutorial, we will use several auxiliary functions. The most important is the run_example core function, which generates a response from the Florence-2 model.

Additionally, we utilize auxiliary functions to visualize the results (draw_bbox ,draw_ocr_bboxes and draw_polygon) and handle the conversion between bounding boxes formats (convert_bbox_to_florence-2 and convert_florence-2_to_bbox). These can be explored in the attached Colab notebook.

In this tutorial we introduce Florence-2 [1]— a novel, open-source Vision-Language Model (VLM) designed to handle a diverse range of vision and multimodal tasks, including captioning, object detection, segmentation and OCR.

In FT-IR spectroscopy, precise light guidance from source to sensor is crucial. Our off-axis parabolic mirrors are designed to direct light efficiently, minimizing loss and enhancing the quality of the spectral data. These mirrors are crafted with high precision to ensure that every photon is utilized, resulting in superior sensitivity and accuracy in your measurements.

Inspired by Large Language Models (LLMs), Florence-2 was designed as a sequence-to-sequence model. It takes an image and text instructions as inputs, and outputs text results. The input or output text may represent plain text or a region in the image. The region format varies depending on the task:

mobilevlm v2: faster and stronger baseline for vision language model

For your convenience, I added a widget to the Colab notebook that enables you to draw a bounding box on the image, and code to convert it to Florence-2 format.

Our windows and window coatings are designed to withstand environmental conditions such as temperature changes, humidity, abrasion, and mechanical stress, ensuring that the performance of the optical system remains unaffected. Optical windows can be made from various materials, such as glass, fused silica, or crystals, depending on the specific requirements of the application, such as the wavelength range and environmental durability needed.

After installing and importing the necessary libraries (as demonstrated in the accompanying Colab notebook), we begin by loading the Florence-2 model, processor and the input image of a camera:

Diffraction gratings are essential components in dispersing light into its constituent wavelengths. In FT-IR spectroscopy, they allow for the separation of complex infrared signals into detailed spectra. Our diffraction gratings are engineered to provide high efficiency and resolution, ensuring that even the most subtle spectral features are discernible. This precision enables researchers to uncover intricate details about their samples.

Florence-2 can perform a variety of visual tasks. Let’s explore some of its capabilities, starting with image captioning.

Omega coating heritage dates back to 1936, and spans a wide range of technologies: Sunglasses to space-borne sun-visors; halogen light bulbs to COVID-fighting UV-C lamps, supermarket scanners to modern printers, astronomical mirrors to prestigious art installations to entertainment optics: if an application requires a precision or decorative coatings we surely have a proven solution.

Our commitment to finding the right solutions for difficult problems means that we are capable of doing what other engineers cannot. Some say we bend physics and solve the unsolvable. Can we? Maybe. What we know we can do is this: We ask questions and we dig, because sometimes the solutions people believe they need only scratch the surface

Omega Optical also provides a range of FT-IR accessory optics, including Attenuated Total Reflection (ATR) sampling optics and liquid cells.

The primary function of a beamsplitter is to separate a single beam of light into two parts, one reflected and one transmitted. The incident light is split at the surface which is usually set at some angle so the reflected and transmitted beams are separated. Any metal layer that forms a partially reflective mirror can serve as a beamsplitter, and many under the “plate” category are no more sophisticated than that.

Lastly, the model’s output is decoded and post-processed with processor.batch_decode and processor.post_process_generation to produce the final text response, which is returned by the run_example function.

Discover how Omega Optical can enhance your FT-IR spectroscopy capabilities. Contact us today to learn more about our products and services.

We are capable to perform every lens and mirror shaping and polishing step after material synthesis. A multitude of conventional spindles combined with CNC grinding and diamond turning can fabricate virtually any curved optic.

Optical filters play a crucial role in FT-IR spectroscopy by selectively transmitting desired wavelengths and blocking out unwanted ones. Our filters are designed to optimize signal-to-noise ratios, ensuring that only relevant spectral information reaches the detector. With a range of custom filters tailored for specific applications, Omega Optical helps you achieve the highest level of precision and accuracy in your FT-IR measurements.